KR20240010457A - Systems, methods, kits and devices for digital product network systems and biology-based value chain networks - Google Patents

Abstract

A digital product network system generally includes a set of digital products each having a product memory, a product network interface, and a product processor programmed with product instructions; A product network control tower having a control tower memory, a control tower network interface, and a control tower processor programmed with control tower instructions; and a digital twin system defined at least in part by at least one of product instructions or control tower instructions to encode a set of digital twins representing a set of digital products.

Description

Systems, methods, kits and devices for digital product network systems and biology-based value chain networks

Cross-reference to related applications

This application is filed under U.S. Provisional Patent Application No. 63/176,198, filed on April 16, 2021, and U.S. Provisional Patent Application No. 63/282,507, filed on November 23, 2021, on January 14, 2022. Claims the benefit of U.S. Provisional Patent Application No. 63/299,710, filed on January 21, 2022, and U.S. Provisional Patent Application No. 63/302,013, filed on January 21, 2022. This application claims priority to Indian patent application IN202211008709 filed on February 18, 2022. The entire disclosures of the above applications are incorporated herein by reference.

Field

This disclosure relates to information technology methods and systems for management of value chain network entities, including supply chain and demand management entities. The present disclosure also relates to the field of enterprise management platforms, and more specifically involves quantum and biology-based value chain networks.

Historically, the majority of the various categories of goods purchased and used by household consumers, businesses, and other customers were supplied primarily through relatively linear arrangements, in which manufacturers and other suppliers of finished goods, components, and other items The item is handed over to a delivery company, freight forwarder, etc., and the delivery company, freight forwarder, etc. deliver the item to a warehouse for temporary storage, to a retailer where the customer purchases the item, or directly to the customer's location. Manufacturers and retailers perform a variety of sales and marketing activities to encourage and satisfy customer demand, including designing products, positioning and advertising products on shelves, and setting prices. did.

Orders for products are processed by manufacturers through supply chains such as those shown in FIG. 1 and various supply environments 160, which operate production facilities 134 or act as resellers or distributors for others. A supplier 122 at responds to the order and obtains product 130 at the point of origin 102. Product 130 is delivered through the supply chain and is connected to various transportation facilities 138 and distribution facilities 134, such as warehouses 132, fulfillment centers 112, and delivery systems 114; For example, it is transported and stored by trucks, other vehicles, trains, etc. In many cases, transportation by waterway is provided between a point of origin 102 and one or more destinations 104 via maritime facilities and infrastructure such as ships, barges, docks, and ports.

Organizations have access to virtually unlimited amounts of data. With the advent of smart connected devices, wearable technology, and the Internet of Things (IoT), the amount of data available to organizations planning, overseeing, managing, and operating value chain networks has increased dramatically and is likely to continue to do so. . For example, in a manufacturing facility, warehouse, campus, or other operational environment, vibration data to measure the vibration signature of critical machines, motion sensors to track temperature and throughput throughout the facility, and locating items. There may be hundreds or thousands of IoT sensors providing metrics such as asset tracking sensors and beacons, cameras and optical sensors, chemical and biological sensors, etc. Additionally, as wearable technology becomes more prevalent, wearables can provide insight into workers' movements, health indicators, physiological state, activity status, movement, and other characteristics. Additionally, as organizations implement CRM systems, ERP systems, operational systems, information technology systems, advanced analytics, and other systems that leverage information and information technology, the data sets and Access third-party data sets and an increasingly wide range of other large-scale data sets, such as marketing data, sales data, operational data, information technology data, performance data, customer data, financial data, market data, pricing data, supply chain data, and more. You can.

Although the existence of more data and new types of data provides many opportunities for organizations to achieve competitive advantage; However, this also raises issues such as complexity and capacity, which can lead to users being overwhelmed and missing opportunities for insight. There is a need for methods and systems that enable businesses to not only acquire data, but also transform data into insights, and insights into timely execution of well-informed decisions and efficient operations.

In value chain networks, it has become more common to acquire large data sets from thousands or potentially millions of devices (including large numbers of sensors) distributed across multiple organizations. For example, RFID (Radio Frequency Identification) tags are proliferating on individual products in retail stores. In this and other similar situations, the vast number of data streams can overwhelm the ability to transmit data over networks and/or generate effective automated centralized decisions.

The proliferation of data generators (e.g., sensors) has created an opportunity to manage networks such as value chain networks with input from a vast number of distributed points of semi-intelligent control. However, current approaches rely on centralized data collection that is often limited due to bandwidth, storage, processing, and/or other limitations.

One aspect of the present disclosure relates to a method for transmitting a predictive model of a data stream from a first device to a second device. The method may include receiving, by a first device, a plurality of data values of a data stream. Data values may include sensor data collected from one or more sensor devices. The method may include generating, by the first device, a prediction model to predict future data values of the data stream based on the plurality of data values received. Creating a prediction model may include determining a plurality of model parameters. The method may include transmitting, by the first device, a plurality of model parameters to the second device. The method may include receiving, by a second device, a plurality of model parameters. The method may include parameterizing, by a second device, a prediction model using a plurality of model parameters. The method may include predicting, by the second device, future data values of the data stream using a parameterized prediction model. In embodiments, parameters include vectors. In embodiments, the vector is a motion vector associated with the robot. In embodiments, the future data values of the data stream include one or more future predicted positions of the robot. In embodiments, the predictive model predicts a stock level of an item, and the method further includes detecting an upcoming shortage of the item based on future data values. The method may further include taking steps to avoid shortage of the item. In embodiments, the prediction model is a behavior analysis model. In embodiments, future data values represent predicted behavior of an entity. In embodiments, the predictive model is an augmented model, where future data values correspond to an inoperative sensor. In embodiments, the prediction model is a classification model. In embodiments, future data values represent a predicted future state of a system including one or more sensor devices. In embodiments, the sensors are RFID sensors associated with cargo. In embodiments, future data values represent future locations of the cargo. In embodiments, the sensors are security cameras. In embodiments, the data stream includes motion vectors extracted from video data captured by security cameras. In embodiments, the sensors are vibration sensors that measure vibrations produced by machines. In embodiments, future data values indicate a potential need for maintenance of machines. The method may further include receiving, by the first device, additional data values of the data stream. The method may include refining, by the first device, the prediction model using additional data values. In embodiments, refining the predictive model adjusts model parameters. The method may include transmitting the adjusted model parameters to the second device. The method may further include receiving, by the second device, the adjusted model parameters. The method may include re-parameterizing the prediction model using the adjusted model parameters. The method may include generating additional future data values using the re-parameterized prediction model.

Another aspect of the present disclosure relates to a method for prioritizing predictive model data streams. The method may include receiving, by a first device, a plurality of predictive model data streams. In embodiments, each of the prediction model data streams includes a set of model parameters for the corresponding prediction model. In embodiments, each prediction model is trained to predict future data values of a data source. The method may include prioritizing, by the first device, priorities for each of the plurality of predictive model data streams. The method may include selecting at least one of the predictive model data streams based on a corresponding priority. The method may include parameterizing, by the first device, a prediction model using a set of model parameters included in the selected prediction model stream. The method may include predicting, by the first device, future data values of a data source using a parameterized prediction model. In embodiments, the selected at least one predictive model data stream is associated with high priority. In embodiments, selecting includes suppressing unselected predictive model data streams based on priorities associated with each unselected predictive model data stream. In embodiments, assigning priorities to each of the plurality of predictive model data streams includes determining whether each set of model parameters is abnormal. In embodiments, assigning priorities to each of the plurality of predictive model data streams includes determining whether each set of model parameters has changed from a previous value. In embodiments, the set of model parameters includes at least one vector. In an embodiment, the at least one vector includes a motion vector associated with the robot. In embodiments, the future data values include one or more future predicted positions of the robot. In some embodiments, the predictive model predicts stock levels of items. The method may further include detecting an upcoming shortage of an item based on future data values. The method may include taking steps to avoid shortage of the item. In embodiments, the predictive model is a behavior analysis model. In embodiments, future data values represent predicted behavior of an entity. In some embodiments, the predictive model is an augmented model. In embodiments, future data values correspond to a non-operational sensor. In embodiments, the prediction model is a classification model. In some embodiments, future data values represent a predicted future state of a system including one or more sensor devices. In embodiments, the sensors are RFID sensors associated with the cargo. In some embodiments, the future data values represent the future location of the cargo. In embodiments, the sensors are security cameras. In some embodiments, the data stream includes motion vectors extracted from video data captured by security cameras. In embodiments, the sensors are vibration sensors that measure vibrations produced by machines. In some embodiments, future data values indicate a potential need for maintenance of machines.

In accordance with some embodiments of the present disclosure, a computer-implemented method is disclosed substantially as described above with reference to any of the examples and/or any of the accompanying drawings. According to some embodiments of the present disclosure, a computing system comprising one or more processors and one or more memory configured to perform operations substantially as described above with reference to any of the examples and/or any of the accompanying drawings. This begins. According to some embodiments of the present disclosure, when executed across one or more processors, at least a portion of the one or more processors performs operations substantially as described above with reference to any of the examples and/or any of the accompanying drawings. A computer program product residing on a computer-readable storage medium storing a plurality of instructions for performing instructions is disclosed. According to some embodiments of the present disclosure, a device configured substantially as described above with reference to any of the examples and/or any of the accompanying drawings is disclosed.

A more complete understanding of the present disclosure will be gained from the following description and accompanying drawings and claims. All documents referenced in this application are incorporated herein by reference.

The accompanying drawings, which are included to provide a better understanding of the disclosure, illustrate embodiments of the disclosure and, together with the description, serve to explain many aspects of the disclosure. In the drawing:
1 is a block diagram illustrating the prior art relationships of various entities and facilities in a supply chain.
Figure 2 is a block diagram showing the components and interrelationships of the system and process of the value chain network according to the present disclosure.
3 is another block diagram illustrating the components and interrelationships of the system and process of the value chain network according to the present disclosure.
FIG. 4 is a block diagram illustrating the components and interrelationships of the systems and processes of the digital product network of FIGS. 2 and 3 according to the present disclosure.
Figure 5 is a block diagram showing the components and interrelationships of the system and process of the value chain network technology stack according to the present disclosure.
FIG. 6 is a block diagram illustrating a platform and relationships for orchestrating control of various entities in a value chain network according to the present disclosure.
Figure 7 is a block diagram showing components and relationships in an embodiment of a value chain network management platform according to the present disclosure.
8 is a block diagram illustrating components and relationships of value chain entities managed by an embodiment of the value chain network management platform according to the present disclosure.
Figure 9 is a block diagram showing the network relationship of entities in the value chain network according to the present disclosure.
10 is a block diagram illustrating a set of applications supported by the integrated data handling layer in the value chain network management platform according to the present disclosure.
11 is a block diagram illustrating components and relationships in an embodiment of a value chain network management platform according to the present disclosure.
Figure 12 is a block diagram showing the components and relationships of the data storage layer in an embodiment of the value chain network management platform according to the present disclosure.
Figure 13 is a block diagram showing the components and relationships of the adaptive intelligence system layer in an embodiment of the value chain network management platform according to the present disclosure.
14 is a block diagram illustrating providing an adaptive intelligence system for coordinated intelligence for a set of supply and demand applications for a category of goods in accordance with the present disclosure.
FIG. 15 is a block diagram illustrating providing a hybrid adaptive intelligence system for coordinated intelligence for a set of demand and supply applications or categories of goods in accordance with the present disclosure.
FIG. 16 is a block diagram illustrating providing an adaptive intelligence system for predictive intelligence for a set of supply and demand applications for a category of goods in accordance with the present disclosure.
FIG. 17 is a block diagram illustrating providing an adaptive intelligence system for classification intelligence for a set of supply and demand applications for categories of products in accordance with the present disclosure.
FIG. 18 is a block diagram illustrating providing an adaptive intelligence system for generating automated control signals for a set of supply and demand applications for a category of goods in accordance with the present disclosure.
19 is a block diagram illustrating training an artificial intelligence/machine learning system to generate information routing recommendations for a selected value chain network in accordance with this disclosure.
Figure 20 is a block diagram illustrating a semi-perceptual problem recognition system for recognition of pain points/problem states in a value chain network according to the present disclosure.
21 is a block diagram illustrating a set of artificial intelligence systems operating on value chain information to enable automated coordination of value chain activities for an enterprise according to the present disclosure.
FIG. 22 is a block diagram illustrating the components and relationships involved in integrating a set of digital twins in an embodiment of a value chain network management platform according to the present disclosure.
FIG. 23 is a block diagram illustrating a set of digital twins involved in an embodiment of a value chain network management platform according to the present disclosure.
FIG. 24 is a block diagram illustrating components and relationships of an entity discovery and management system in an embodiment of a value chain network management platform according to the present disclosure.
Figure 25 is a block diagram showing components and relationships of a robotic process automation system in an embodiment of the value chain network management platform according to the present disclosure.
FIG. 26 is a block diagram illustrating the components and relationships of a set of opportunity miners in an embodiment of a value chain network management platform according to the present disclosure.
FIG. 27 is a block diagram illustrating the components and relationships of a set of edge intelligence systems in an embodiment of a value chain network management platform according to the present disclosure.
Figure 28 is a block diagram showing components and relationships in an embodiment of a value chain network management platform according to the present disclosure.
FIG. 29 is a block diagram illustrating additional details of components and relationships in an embodiment of a value chain network management platform according to the present disclosure.
FIG. 30 is a block diagram illustrating components and relationships in an embodiment of a value chain network management platform enabling centralized orchestration of value chain network entities according to this disclosure.
Figure 31 is a block diagram showing the components and relationships of an integrated database in an embodiment of the value chain network management platform according to the present disclosure.
32 is a block diagram illustrating the components and relationships of a set of integrated data collection systems in an embodiment of a value chain network management platform according to the present disclosure.
Figure 33 is a block diagram showing the components and relationships of a set of Internet of Things monitoring systems in an embodiment of the value chain network management platform according to the present disclosure.
34 is a block diagram illustrating the components and relationships of a machine vision system and a digital twin in an embodiment of a value chain network management platform according to the present disclosure.
35 is a block diagram illustrating the components and relationships of a set of adaptive edge intelligence systems in an embodiment of a value chain network management platform according to the present disclosure.
36 is a block diagram illustrating additional details of the components and relationships of a set of adaptive edge intelligence systems in an embodiment of a value chain network management platform according to the present disclosure.
Figure 37 is a block diagram illustrating the components and relationships of a set of integrated adaptive intelligence systems in an embodiment of a value chain network management platform according to the present disclosure.
38 is a schematic diagram of a system configured to train artificial systems utilized by value chain systems and digital twin systems using real-world outcome data, according to some embodiments of the present disclosure.
39 is a schematic diagram of a system configured to train artificial systems utilized by a container fleet management system and digital twin system using real-world outcome data, according to some embodiments of the present disclosure.
40 is a schematic diagram of a system configured to train artificial systems utilized by a logistics design system and digital twin system using real-world outcome data, according to some embodiments of the present disclosure.
Figure 41 is a schematic diagram of a system configured to train artificial systems utilized by a packaging design system and digital twin system using real-world results data, according to some embodiments of the present disclosure.
42 is a schematic diagram of a system configured to train artificial systems utilized by waste mitigation systems and digital twin systems using real-world outcome data, according to some embodiments of the present disclosure.
43 is a schematic diagram illustrating an example of a portion of an information technology system for value chain artificial intelligence utilizing digital twins according to some embodiments of the present disclosure.
44 is a block diagram illustrating the components and relationships of a set of intelligent project management facilities in an embodiment of a value chain network management platform according to the present disclosure.
Figure 45 is a block diagram showing the components and relationships of an intelligent task recommendation system in an embodiment of the value chain network management platform according to the present disclosure.
Figure 46 is a block diagram showing the components and relationships of a routing system between nodes of a value chain network in an embodiment of the value chain network management platform according to the present disclosure.
Figure 47 is a block diagram illustrating the components and relationships of a dashboard for managing a set of digital twins in an embodiment of a value chain network management platform.
Figure 48 is a block diagram illustrating components and relationships in an embodiment of a value chain network management platform using a microservices architecture.
Figure 49 is a block diagram illustrating the components and relationships of an Internet of Things data collection architecture and sensor recommendation system in an embodiment of a value chain network management platform.
Figure 50 is a block diagram illustrating the components and relationships of a social data collection architecture in an embodiment of a value chain network management platform.
Figure 51 is a block diagram illustrating the components and relationships of a crowdsourcing data collection architecture in an embodiment of a value chain network management platform.
Figure 52 is a schematic diagram illustrating an embodiment of a set of value chain network digital twins representing a virtual model of a set of value chain network entities according to the present disclosure.
Figure 53 is a schematic diagram showing an embodiment of a warehouse digital twin kit system according to the present disclosure.
Figure 54 is a schematic diagram showing an embodiment of a stress test performed on a value chain network according to the present disclosure.
Figure 55 is a schematic diagram illustrating an embodiment of a method used by a machine to detect defects and predict any future failure of the machine in accordance with the present disclosure.
Figure 56 is a schematic diagram illustrating an embodiment of the deployment of machine twins to perform predictive maintenance on a set of machines according to the present disclosure.
Figure 57 is a schematic diagram illustrating an example of a portion of a system for value chain customer digital twin and customer profile digital twin according to some embodiments of the present disclosure.
Figure 58 is a schematic diagram illustrating an example of an advertising application interfacing with an adaptive intelligence system layer according to the present disclosure.
Figure 59 is a schematic diagram illustrating an example of an e-commerce application integrated with an adaptive intelligence system layer according to the present disclosure.
Figure 60 is a schematic diagram illustrating an example of a demand management application integrated with an adaptive intelligence system layer according to the present disclosure.
61 is a schematic diagram illustrating an example of a portion of a system for value chain smart supply component digital twin according to some embodiments of the present disclosure.
Figure 62 is a schematic diagram illustrating an example of a risk management application interfacing with an adaptive intelligence system layer according to the present disclosure.
63 is a schematic diagram of maritime assets associated with a value chain network management platform including components of port infrastructure according to the present disclosure.
64 and 65 are schematic diagrams of maritime assets associated with a value chain network management platform including components of a vessel according to the present disclosure.
66 is a schematic diagram of maritime assets associated with a value chain network management platform including components of a barge according to the present disclosure.
67 is a schematic diagram of maritime assets associated with a value chain network management platform, including those involving maritime events, legal proceedings, and utilizing geofenced parameters in accordance with this disclosure.
Figure 68 is a schematic diagram illustrating an example environment of an enterprise and executive control tower and management platform, including data sources in communication therewith, according to some embodiments of the present disclosure.
Figure 69 is a schematic diagram illustrating an example set of components of an enterprise control tower and management platform according to some embodiments of the present disclosure.
Figure 70 is a schematic diagram illustrating an example of an enterprise data model according to some embodiments of the present disclosure.
Figure 71 is a schematic diagram illustrating examples of different types of enterprise digital twins, including executive digital twins, in relation to the data layer, processing layer, and application layer of an enterprise digital twin framework according to some embodiments of the present disclosure.
Figure 72 is a schematic diagram illustrating an example implementation of an enterprise and executive control tower and management platform in accordance with some embodiments of the present disclosure.
Figure 73 is a flow diagram illustrating an example set of operations for configuring and serving an enterprise digital twin.
Figure 74 illustrates an example set of operations of a method for constructing an organizational digital twin.
Figure 75 illustrates an example set of operations of a method for creating an executive digital twin.
76-103 include those involving expert systems, self-organization, machine learning, artificial intelligence, for pattern recognition, for classification of one or more parameters, characteristics, or phenomena, according to embodiments of the present disclosure; A schematic diagram of an embodiment of a neural network system that may be connected to, integrated with, and accessible by the platform to enable intelligent transactions, including trained neural network systems, for support of autonomous control, and for other purposes. .
104 is a schematic diagram illustrating an example intelligent service system according to some embodiments of the present disclosure.
105 is a schematic diagram illustrating an example neural network with multiple layers according to some embodiments of the present disclosure.
106 is a schematic diagram illustrating an example convolutional neural network (CNN) according to some embodiments of the present disclosure.
107 is a schematic diagram illustrating an example neural network for implementing natural language processing according to some embodiments of the present disclosure.
108 is a schematic diagram illustrating an example reinforcement learning-based approach for executing one or more tasks by a mobile system in accordance with some embodiments of the present disclosure.
109 is a schematic diagram illustrating an example physical orientation crystal chip according to some embodiments of the present disclosure.
110 is a schematic diagram illustrating an example network enhancement chip according to some embodiments of the present disclosure.
Figure 111 is a schematic diagram illustrating an example diagnostic chip according to some embodiments of the present disclosure.
112 is a schematic diagram illustrating an example governance chip according to some embodiments of the present disclosure.
Figure 113 is a schematic diagram illustrating an example prediction, classification, and recommendation chip according to some embodiments of the present disclosure.
Figure 114 is a schematic diagram illustrating an example environment of an autonomous additive manufacturing platform according to some embodiments of the present disclosure.
Figure 115 is a schematic diagram illustrating an example implementation of an autonomous additive manufacturing platform for automating and optimizing digital production workflows for metal additive manufacturing, according to some embodiments of the present disclosure.
Figure 116 is a flow chart illustrating optimization of different parameters of an additive manufacturing process according to some embodiments of the present disclosure.
117 is a schematic diagram illustrating a system for learning based on data from an autonomous additive manufacturing platform to train an artificial learning system to use a digital twin for classification, prediction, and decision-making, according to some embodiments of the present disclosure. am.
118 is a schematic diagram illustrating an example implementation of an autonomous additive manufacturing platform including various components along with other entities in a distributed manufacturing network according to some embodiments of the present disclosure.
119 automates manufacturing functions and subprocesses, including process and material selection, hybrid part workflow, feedstock composition, part design optimization, risk prediction and management, marketing and customer service, according to some embodiments of the present disclosure; Schematic diagram illustrating an example implementation of an autonomous additive manufacturing platform for managing.
120 is a schematic diagram of a distributed manufacturing network built on a distributed ledger system and enabled by an autonomous additive manufacturing platform, according to some embodiments of the present disclosure.
121 illustrates an example implementation of a distributed manufacturing network where digital thread data is tokenized and stored in a distributed ledger to ensure traceability of printed parts across one or more manufacturing nodes of the distributed manufacturing network according to some embodiments of the present disclosure. This is a schematic diagram illustrating.
122 is a schematic diagram illustrating an example implementation of a conventional computer vision system for generating images of objects of interest.
Figure 123 is a schematic diagram illustrating an example implementation of a dynamic vision system for dynamically learning object concepts regarding an object of interest according to some embodiments of the present disclosure.
124 is a schematic diagram illustrating an example architecture of a dynamic vision system according to some embodiments of the present disclosure.
125 is a flowchart illustrating a method for object recognition by a dynamic vision system according to some embodiments of the present disclosure.
126 is a schematic diagram illustrating an example implementation of a dynamic vision system for modeling, simulating, and optimizing various optical, mechanical, design, and lighting parameters of a dynamic vision system in accordance with some embodiments of the present disclosure.
127 is a schematic diagram illustrating an example implementation of a dynamic vision system displaying details of various components along with integration of the dynamic vision system with one or more third-party systems, according to some embodiments of the present disclosure.
128 is a schematic diagram illustrating an example environment of a fleet management platform according to some embodiments of the present disclosure.
Figure 129 is a schematic diagram illustrating example configurations of a multipurpose robot and a special purpose robot according to some embodiments of the present disclosure.
130 is a schematic diagram illustrating an example platform-level intelligence layer of a fleet management platform according to some embodiments of the present disclosure.
Figure 131 is a schematic diagram illustrating an example configuration of an intelligence layer according to some embodiments of the present disclosure.
Figure 132 is a schematic diagram illustrating an example security framework according to some embodiments of the present disclosure.
Figure 133 is a schematic diagram illustrating an example environment of a fleet management platform according to some embodiments of the present disclosure.
Figure 134 is a schematic diagram illustrating example data flow of a task organization system according to some embodiments of the present disclosure.
Figure 135 is a schematic diagram illustrating example data flow of a flit operation system according to some embodiments of the present disclosure.
136 is a schematic diagram illustrating an example task parsing system and task definition system and example data flows thereof according to some embodiments of the present disclosure.
Figure 137 is a schematic diagram illustrating an example fleet configuration system and example data flows thereof according to some embodiments of the present disclosure.
138 is a schematic diagram illustrating an example workflow definition system and example data flows according to some embodiments of the present disclosure.
Figure 139 is a schematic diagram illustrating an example configuration of a multipurpose robot and its components according to some embodiments of the present disclosure.
140 is a schematic diagram illustrating an example architecture of a robot control system according to some embodiments of the present disclosure.
141 is a schematic diagram illustrating an example architecture of a robot control system 12150 utilizing data from multiple sensors in a vision and sensing system in accordance with some embodiments of the present disclosure.
142 is a schematic diagram illustrating an example vision and sensing system for a robot according to some embodiments of the present disclosure.
Figure 143 is a schematic diagram illustrating an example process executed by a multi-purpose robot to harvest crops, according to some embodiments of the present disclosure.
Figure 144 is a schematic diagram illustrating an example environment of an intermodal smart container system according to some embodiments of the present disclosure.
Figure 145 is a schematic diagram illustrating an example configuration of a smart container according to some embodiments of the present disclosure.
Figure 146 is a schematic diagram illustrating an intelligent service adapted to provide intelligent services in a smart intermodal container system according to some embodiments of the present disclosure.
Figure 147 is a schematic diagram illustrating a digital twin module according to some embodiments of the present disclosure.
Figure 148 illustrates an example embodiment of a method for receiving a request to update one or more characteristics of a digital twin of a shipping entity and/or environment.
Figure 149 illustrates an example embodiment of a method for updating a set of cost of downtime values in a digital twin of a smart container according to some embodiments of the present disclosure.
Figure 150 is a schematic diagram illustrating an example environment of a digital product network according to some embodiments of the present disclosure.
Figure 151 is a schematic diagram illustrating an example environment of a connected product according to some embodiments of the present disclosure.
Figure 152 is a schematic diagram illustrating an example environment of a digital product network according to some embodiments of the present disclosure.
Figure 153 is a schematic diagram illustrating an example environment of a digital product network according to some embodiments of the present disclosure.
Figure 154 is a flow chart illustrating a method of using product level data according to some embodiments of the present disclosure.
Figure 155 is a schematic diagram illustrating an example environment of a digital product network according to some embodiments of the present disclosure.
Figure 156 is a schematic diagram illustrating an example of a smart futures contract system according to some embodiments of the present disclosure.
Figure 157 is a schematic diagram illustrating an example environment of an edge networking system according to some embodiments of the present disclosure.
Figure 158 is a schematic diagram illustrating an example environment of an edge networking system including a VCN bus according to some embodiments of the present disclosure.
Figure 159 is a schematic diagram illustrating an example environment of an edge networking system according to some embodiments of the present disclosure, including a configured device EDNW system.
160 is a schematic diagram of an example embodiment of a quantum computing service according to some embodiments of the present disclosure.
Figure 161 illustrates quantum computing service request handling according to some embodiments of the present disclosure.
Figure 162 is a schematic diagram illustrating embodiments of a biology-based value chain network system according to the present disclosure.
Figure 163 is a schematic diagram of the thalamus service and how it is coordinated within modules according to the present disclosure.
Figure 164 is a block diagram illustrating an energy system that can communicate with similar systems, subsystems, components, and a value chain network management platform in accordance with some embodiments of the present disclosure.
Figure 165 is a block diagram illustrating a schematic diagram of a dual-process artificial neural network system according to some embodiments of the present disclosure.

Over time, the company developed software systems to predict and manage customer demand, RFID and asset tracking systems to track goods as they move through the supply chain, and navigation and routing to improve the efficiency of route selection. Technological solutions, such as systems, are increasingly being used to improve the outcomes associated with traditional supply chains, such as those shown in Figure 1. However, several larger trends are placing increasing pressure on manufacturers, retailers and other operators to improve supply chain performance. First, online and e-commerce operators, especially Amazon™, have become the largest retail channel for goods in many categories, and in some cases the day after a customer orders an item, and in some cases the same day (and in some cases, drones, robots, and/or delivered to your doorstep by autonomous vehicles) have introduced distribution and fulfillment centers (112) throughout some regions, such as the United States, housing hundreds of thousands, and sometimes more, of product categories (SKUs). For retailers without a wide geographic distribution of fulfillment centers or warehouses, customer expectations for speed of delivery have placed increased pressure on supply chain efficiency and optimization. Accordingly, there still exists a need for improved supply chain methods and systems.

Second, agile manufacturing capabilities (e.g., using 3D printing and robotic assembly technologies, among others), customer profiling technologies, and online ratings and reviews have increased customer expectations for customization and personalization of products. Accordingly, to compete, manufacturers and retailers need improved methods and systems for understanding, predicting, and satisfying customer demand.

Historically, supply chain management and demand planning and management were mostly separate activities that were integrated primarily when demand was converted into orders, which were then passed on to the supply side for processing in the supply chain. As expectations for speed and personalization increase, there is a need for methods and systems that can provide integrated orchestration of supply and demand.

Parallel to these other big trends, the Internet of Things has emerged, and some categories of products, especially smart home products such as thermostats, lighting systems, and speakers, can provide specific functions such as playing music or even ordering products. They are becoming increasingly capable with onboard network connectivity and processing capabilities, often including voice-controlled intelligent agents such as Alexa™ or Siri™ that allow device control and triggering of application features. In some cases, smart products 650, such as printers that order refill cartridges, may even initiate the order. Intelligent products 650 may, in some cases, be used as coordinated systems, such as when an Amazon™ Echo™ product controls a television, or when a sensor-enabled thermostat or security camera connects to a mobile device. However, most intelligent products still primarily engage in a set of isolated, application-specific interactions. As artificial intelligence capabilities increase, and the supply environment 670, the demand environment 672, and the product from the manufacturer's loading dock to the destination point 612 of the customer 662 or retailer 664 (1510) As more and more computing and networking power moves to network-enabled edge devices and systems residing in every location, system, and facility that fills the path, dramatic improvements in all factors involved in supply and demand. There is a need and opportunity for advanced intelligence, control, and automation.

VALUE CHAIN NETWORKS

2, a block diagram illustrating the components and interrelationships of the systems and processes of a value chain network is presented at 200. In an exemplary embodiment, a “value chain network,” as used in this application, refers to historically separate demand management systems and processes and supply chain management, enabled by the development and convergence of a number of different technologies. Refers to the elements and interconnections of systems and processes. In an exemplary embodiment, value chain control tower 260 (e.g., referred to herein in some cases as a “value chain network management platform,” “VCNP,” or simply “system,” or “platform”) may be connected to, in communication with, or otherwise operably coupled to a data processing facility, which may be, for example, a customer (e.g., directly connected customer 250), or A big data center that receives data flows, data pools, data streams and/or other data configurations and transmission modalities directly from some other third party 220, for example, from digital product network 21002. (e.g., big data processing 230) and related processing functionality. As described in this application, demand enhancement (262), synchronized planning (234), intelligent procurement (238), dynamic order processing (240), or some other smart action informed by coordinated and adaptive intelligence. For purposes such as market orchestration activities and communications 210, analytics 232, or communications related to some other type of input, communications may also be utilized by the value chain control tower.

3, another block diagram is presented illustrating the components and interrelationships of the systems and processes of a value chain network, and associated use cases, data handling, and associated entities. In an exemplary embodiment, value chain control tower 360 performs demand curve management 352, synchronization of ecosystem 348, intelligent procurement 344, dynamic order processing 350, value chain analysis 340, and/ or may coordinate market orchestration activities 310, including but not limited to smart supply chain operations 342. In an exemplary embodiment, value chain control tower 360 may include an intelligent, user-adaptive interface, adaptive intelligence and control 332, and/or adaptive data monitoring and storage 334, as described herein. ) is further connected to, in communication with, or otherwise operably coupled to an external data source 320 and a data handling stack 330 (e.g., a value chain network technology), which may include It may be connected to, in communication with, or otherwise operably coupled with the adaptive data pipeline 302 and processing facility. Value chain control tower 302 also includes, but is not limited to, digital product network 21002, customers (e.g., direct connect customers 362), and/or other connected operations 364 and entities in the value chain network. It may be further connected to, communicate with, or otherwise operably coupled with additional value chain entities.

DIGITAL PRODUCT NETWORKS (“DPN”)

Referring to Figure 4, a block diagram is presented at 400 illustrating the components and interrelationships of the systems and processes of a digital product network. In example embodiments, products (including goods and services) may generate and transmit data, such as product level data, to a communication layer and/or edge data processing facility within the value chain network technology stack. This data may generate enhanced product level data and may be combined with third party data for further processing, modeling or other adaptive or coordinated intelligence activities, as described herein. This may include, but is not limited to, producing and/or simulating product and value chain use cases, and data therefor may be utilized by the product, product development process, product design, etc.

STACK VIEW EXAMPLES

5, a presentation layer, an intelligence layer, and platforms (e.g., development and hosting platforms), data facilities (e.g., related to data with IoT and big data), and data aggregation facilities. A block diagram illustrating the components and interrelationships of the systems and processes of the value chain network technology stack, which may include, but is not limited to, serverless functionality, is presented at 500. In an example embodiment, the presentation layer may include, but is not limited to, a user interface and modules for research and discovery and tracking the user's experience and engagement. In an example embodiment, the intelligence layer may include, but is not limited to, statistical and computational methods, semantic models, analysis libraries, development environments for analysis, algorithms, logic and rules, and machine learning. In an example embodiment, the platform or value chain network technology stack may include a development environment, APIs for connectivity, cloud and/or hosted applications, and device discovery. In an example embodiment, the data aggregation facility or layer may include, but is not limited to, modules for data normalization for general transmission and heterogeneous data collection from heterogeneous devices. In example embodiments, data facilities or tiers may include, but are not limited to, IoT and big data access, control, and collection and alternatives. In example embodiments, the value chain network technology stack may further be associated with additional data sources and/or technology enablers.

VALUE CHAIN ORCHESTRATION FROM A COMMAND PLATFORM

6 illustrates that the value chain network management platform 604 (referred to in some cases in this application as “value chain control tower”, “VCNP”, or simply “system” or “platform”) manages supply and production factors, demand, and Illustrates a connected value chain network 668 that orchestrates various factors involved in planning, monitoring, controlling, and optimizing the various entities and activities involved in the value chain network 668, such as factors, logistics, and distribution factors. Demand factors are understood by an integrated platform 604 for monitoring and managing supply factors and demand factors as well as status information (e.g., quality and status, planning, ordering and confirmation, and/or tracking and tracing). When orders are created and processed, and when products are created and moved through the supply chain, various entities (e.g., customers/consumers, distributors, suppliers, and producers) (including production, such as production facilities) and may be shared between them. Value chain network 668 may include all of the intelligent products 1510 as well as the equipment, infrastructure, personnel, and other entities involved in planning and satisfying demand for them.

VALUE CHAIN NETWORK AND VALUE CHAIN NETWORK MANAGEMENT PLATFORM

7, the value chain network 668, managed by the value chain management platform 604, may include a set of value chain network entities 652, such as, without limitation: Intelligent Products ( a product 1510 which may be 1510); A set of production facilities 674 involved in producing finished products, components, systems, subsystems, materials used in goods, etc.; various entities, activities and other supply factors 648 involved in the supply environment 670, such as suppliers 642, points of origin 610, etc.; Customers 662 (including consumers, business owners, and intermediate customers such as value-added resellers and distributors), Retailers 664 (online retailers, mobile retailers) located and/or operating in various destinations 612 Now, the various entities, activities and other demand factors 644 involved in the demand environment 672, such as traditional bricks and mortar retailers (including bricks and mortar retailers, pop-up shops, etc.); various distribution environments 678 and distribution facilities 658, such as warehousing facilities 654, fulfillment facilities 628, and delivery systems 632, etc., as well as maritime facilities 622, such as port infrastructure facilities, among others. (660), floating asset (620), and shipyard (638). In embodiments, the value chain network management platform 604 provides a comprehensive set of value chain network 668 processes, workflows, activities, events, and applications 630 (collectively referred to in some cases simply as “applications” 630). capable of monitoring, controlling, and otherwise managing (and in some cases autonomous or semi-autonomous behavior)

Still referring to Figure 7, a high-level schematic diagram of value chain network management platform 604 is illustrated. The value chain network management platform 604 is a network of systems, applications, processes, modules, services, layers, devices, components, machines, products, and services that work cooperatively to enable intelligent management of a set of value chain entities 652. It may include a set of systems, interfaces, connections, and other elements that intelligently manage occurrences, actions, transactions, etc. within one or more value chain network processes, workflows, activities, events and/or applications 630. capable of performing, operating, supporting or enabling, on its own, or otherwise being part of, integrated with, associated with, or operated by VCNP 604 in connection with product 1510. (Products can be finished products, software products, hardware products, component products, materials, equipment items, consumer packaging product items, consumer products, food products, beverage products, household products, business supply products, consumable products, pharmaceutical products, medical It may be a product of any category, such as a device product, a technology product, an entertainment product, or any other type of product and/or set of related services, or, in embodiments, may include an intelligent product 1510, such an intelligent product 1510. Products include, but are not limited to, data processing, networking, sensing, autonomous behavior, intelligent agents, natural language processing, conversation recognition, speech recognition, touch interfaces, remote control, self-organization, self-healing, process automation, computation, and artificial intelligence. , a set of capabilities such as analog or digital sensors, cameras, sound processing systems, data storage, data integration, and/or various Internet of Things capabilities).

In embodiments, management platform 604 may include a set of data handling layers 608, each of which may be used for a wide variety of value chain network applications and end uses, such as automation, mechanical It is structured to provide a set of capabilities that facilitate the development and deployment of intelligence, to facilitate learning, applications of artificial intelligence, intelligent transactions, state management, event management, process management, etc. In an embodiment, the data handling layer 608 is structured in a topology that facilitates the collection and distribution of shared data across multiple applications and uses within the platform 604 by the value chain monitoring system layer 614. Value chain monitoring system layer 614 is for collecting and organizing data collected from or about value chain entities 652, as well as data collected from or about various data layers 624 or services or components. It may include, integrate with, and/or cooperate with various data collection and management systems 640, referred to in some cases for convenience as data collection system 640. In an embodiment, the data handling layer 608 is a platform by value chain network-oriented data storage system layer 624, which in some cases is referred to herein for convenience simply as the data storage layer 624 or storage layer 624. 604 is configured with a topology that facilitates sharing or common data storage across multiple applications and uses. As shown in FIG. 7, data handling layer 608 may also include an adaptive intelligence system layer 614. Adaptive intelligence systems layer 614 may include a set of data processing, artificial intelligence, and computational systems 634 described in greater detail elsewhere throughout this disclosure. Data processing, artificial intelligence, and computational systems 634 may be related to artificial intelligence (e.g., expert systems, artificial intelligence, neural, supervised, machine learning, deep learning, model-based systems, etc.). Specifically, data processing, artificial intelligence, and computational systems 634 may, in some embodiments, utilize recurrent networks, biological systems and the use of opportunity mining (e.g., artificial intelligence systems may be used to monitor new data sources as opportunities to automatically deploy intelligence), robotic process automation (e.g., automation of intelligent agents), edge and network intelligence (e.g., adaptive use of available RF spectrum, adaptive use of available fixed network spectrum, data storage based on available storage conditions) It can be related to various examples such as adaptively storing, coupled to a monitoring system such as adaptively sensing based on some kind of situational sensing, etc.

In embodiments, data handling layer 608 may be represented in the diagram as a vertical stack or ribbon and represents much of the functionality available to platform 604, including storage, monitoring, and processing applications and resources, and combinations thereof. You can. In an embodiment, the set of capabilities of data handling layer 608 may include a shared microservices architecture. In these examples, a set of capabilities may be deployed to provide multiple distinct services or applications, which may be organized as one or more services, workflows, or combinations thereof. In some examples, a set of capabilities may be deployed or reside within a specific application or process. In some examples, a set of capabilities may include one or more activities coordinated for the benefit of the platform. In some examples, a set of capabilities may include one or more events organized for the benefit of the platform. In embodiments, one of the platform's sets of capabilities may be deployed within at least a portion of a common architecture, such as a common architecture that supports a common data schema. In embodiments, one of the sets of capabilities of the platform may be deployed within at least a portion of a common architecture that may support a common repository. In embodiments, one of the sets of capabilities of the platform may be deployed within at least a portion of a common architecture that may support a common monitoring system. In embodiments, one or more sets of capabilities of a platform may be deployed within at least part of a common architecture that may support one or more common processing frameworks. In embodiments, the set of capabilities of data handling layer 608 may include examples where storage capabilities support scalable processing capabilities, scalable monitoring systems, digital twin systems, payment interface systems, etc. In this example, one or more software development kits may be provided by the platform with deployment interfaces to facilitate connection and use of the capabilities of the data handling layer 608. In another example, an adaptive intelligence system may analyze, learn, configure, and reconfigure one or more of the capabilities of data handling layer 608. In embodiments, platform 604 may include a common data storage schema serving, for example, shipyard entity related services and warehousing entity services. There are many other applicable examples and combinations applicable to the foregoing examples that include many of the value chain entities disclosed in this application. In this example, platform 604 may be shown as creating connectivity (e.g., provision of capabilities and information) across many value chain entities. In many examples, pairings of similar kinds of value chain entities that use one or more smaller sets of capabilities of the data handling layer 608 to deploy (interact with, depend on, etc.) a common data schema, common architecture, common interfaces, etc. (Double, triple, quadruple, etc.) exist. Although services and capabilities may be provided to a single value chain entity, a platform can appear to provide countless benefits to the value chain and consumers by supporting connectivity across value chain entities and the applications used by the entities.

VALUE CHAIN NETWORK ENTITIES MANAGED BY THE PLATFORM

8, a value chain network management platform 604 is illustrated in association with a set of value chain entities 652, to which management by the platform 604 may be applied. may be integrated within or with, and/or a broad range of value chain activities (e.g., various value chain network processes, workflows, activities, events and applications 630 (collectively, “Applications 630” or simply “Activities”) "), supply chain activities, logistics activities, demand management and planning activities, delivery activities, delivery activities, warehousing activities, distribution and fulfillment activities, inventory counting, storage and management activities, marketing activities, etc.) It may supply input to and/or take output from the platform 604, such as the accompanying ones. Connections with value chain entities 652 may be facilitated by a set of connection facilities 642 and interfaces 702 that include a wide range of components and systems described throughout this disclosure and in more detail below. This may include, inter alia, connectivity and interface capabilities to individual services of the platform, to the data handling layer, to the platform as a whole, and/or between value chain entities 652.

These value chain entities 652 may be any of the wide variety of assets, systems, devices, machines, components, equipment, facilities, individuals or other entities mentioned throughout this disclosure or in documents incorporated by reference herein; Examples may include, without limitation: machine 724 and its components (e.g., delivery vehicles, forklifts, conveyors, loading machines, cranes, lifts, transporters, trucks, loading machines, unloading machines, packing machines) , picking machines, and robotic systems such as physical robots, collaborative robots (e.g., “cobots”), drones, autonomous vehicles, software bots, and many others); Products (650) (finished products, software products, hardware products, component products, materials, items of equipment, items of consumer packaged goods, consumer products, food products, beverage products, household products, business supply products, consumable products, pharmaceutical products) , may be any category of products, such as medical device products, technology products, entertainment products, or any other type of product and/or set of related services); Value chain process 722 (e.g., shipping process, transport process, maritime process, inspection process, transport process, loading/unloading process, packing/unpacking process, construction process, assembly process, installation process, quality control process, environmental Control processes (e.g. temperature control, humidity control, pressure control, vibration control, etc.), boundary control processes, port-related processes, software processes (including applications, programs, services, etc.), packing and loading processes, financial processes. (e.g. insurance process, reporting process, transaction process, etc.), testing and diagnostic process, security process, safety process, reporting process, asset tracking process, etc.); Wearable and portable devices 720 (e.g., mobile phones, tablets, dedicated portable devices for value chain applications and processes, data collectors (including mobile data collectors), sensor-based devices, watches, glasses, hearing devices, head-mounted devices) devices, clothing-integrated devices, arm bands, bracelets, neck-worn devices, AR/VR devices, headphones, etc.); Workers 718 (e.g., delivery workers, delivery workers, barge workers, dock workers, dock workers, train workers, ship workers, distribution or fulfillment center workers, warehouse workers, vehicle drivers, business managers, engineers, floor managers, Demand managers, marketing managers, inventory managers, supply chain managers, freight handlers, inspectors, delivery personnel, environmental control managers, financial asset managers, process supervisors and operators (for any of the processes mentioned in this application), security personnel, safety personnel, etc.); Suppliers 642 (e.g., suppliers of all types of goods and related services, component suppliers, ingredient suppliers, material suppliers, manufacturers, etc.); Customers 662 (consumers, licensees, businesses, businesses, value-added and other resellers, retailers, end users, distributors, and others who purchase, license, or otherwise use categories of goods and/or related services) (including others who may); Extensive operational facilities 712 (e.g., loading and unloading docks, storage and warehousing facilities 654, vaults, distribution facilities 658, and fulfillment centers 628) (aircraft, airports, hangars, runways, refueling depots, etc.) (including) air travel facilities 740, (piers, yards, cranes, roll-on/roll-off facilities, ramps, containers, container handling systems, waterways 732, locks) Marine facilities 622 (such as port infrastructure facilities 622), shipyard facilities 638, floating assets 620 (such as ships, barges, boats, etc.), point of origin 610 and/or destination. Facilities and other items at points 628, transport facilities 710 (such as container ships, barges, and other floating assets 620), as well as land-based vehicles used to transport goods, such as trucks, trains, etc. vehicles and other delivery systems (632), etc.); Items or factors for which demand is considered (i.e., demand factors 644) (including market factors, events, etc.); Items or factors considered for supply (i.e., supply factors 648) (including market factors, weather, availability of components and materials, etc.); Logistics factors 750 (e.g., availability of travel routes, weather, fuel prices, regulatory factors, availability of space (e.g., on a vehicle, within a container, within a package, within a warehouse, within a fulfillment center, on the shelf) on, etc.), etc.); Retailers 664 (including other retailers such as online retailers 730 and e-commerce sites 730); routes for transportation (e.g., waterways 732, roads 734, air travel routes, railways 738, etc.); Robotic systems 744 (including mobile robots, cobots, robotic systems to assist human workers, robotic delivery systems, etc.); Drones (748) (including those for package delivery, site mapping, monitoring or inspection, etc.); Autonomous vehicles 742 (e.g., for package delivery); Software platform 752 (e.g., enterprise resource planning platform, customer relationship management platform, sales and marketing platform, asset management platform, Internet of Things platform, supply chain management platform, platform as a service platform, infrastructure as a service platform, software-based data storage platform, analysis platform, artificial intelligence platform, etc.); etc. In some example embodiments, product 1510 may be encompassed as intelligent product 1510 or VCNP 604 may include intelligent product 1510 . Intelligent products 1510 may include, but are not limited to, data processing, networking, sensing, autonomous operation, intelligent agents, natural language processing, conversation recognition, speech recognition, touch interfaces, remote control, self-organization, self-healing, and process automation. , may be enabled by a set of capabilities such as computation, artificial intelligence, analog or digital sensors, cameras, sound processing systems, data storage, data integration, and/or various Internet of Things capabilities. Intelligent products 1510 may include some form of information technology. Intelligent product 1510 may have a processor, computer random access memory, and communication modules. Intelligent product 1510 may be a passive intelligent product similar to an RFID type data structure through which the intelligent product can be pinged or read. The product 1510 can be considered a value chain network entity (e.g., under the control of a platform), with the surrounding infrastructure allowing data to be read from the intelligent product 1510, and by adding RFID. It can become intelligent. Intelligent products 1510 may fit into the value chain network in a connected manner such that connectivity is built around intelligent products 1510 through sensors, IoT devices, tags, or other components.

In embodiments, monitoring system layer 614 may monitor any or all of the value chain entities 652 within value chain network 668, exchange data with value chain entities 652, or For example, through various capabilities of data handling layer 608 described throughout this disclosure, it may provide control instructions to or take instructions from any of the value chain entities 652, etc.

NETWORK CHARACTERISTICS OF THE VALUE CHAIN NETWORK ENTITIES

9, the orchestration of a set of deeply interconnected value chain network entities 652 within a value chain network 668 by a value chain network management platform 604 is illustrated. Each of the value chain network entities 652 has a connection to the VCNP 604, a connection to a set of other value chain network entities 652 (local network connections, peer-to-peer connections, mobile may have a network connection, a connection through the cloud, or another connection), and/or a connection to another value chain network entity 652 via VCNP 604. Value chain network management platform 604 may manage the status of entities 652 , e.g., by using information from one set of entities 652 to notify applications 630 involving another set of entities 652 . An edge computation system deployed on or within an entity 652 and its environment, by coordinating the activities of a set of entities 652 or by providing input to an artificial intelligence system of the entities 652 or an artificial intelligence system of the VCNP 604 ( may manage connections, configure or provision resources to enable connectivity, and/or manage applications 630 that utilize connectivity, such as by interacting with an edge computation system.

Entities 652 may be external such that VCNP 604 can interact with these entities 652. When VCNP 604 functions as a control tower to establish monitoring (e.g., to establish monitoring such as common monitoring across several entities 652). In one integrated platform, there may be an interface where the user can view various items such as the user's destination, ports, air and rail assets, as well as orders, etc. At this point, the next step may be to establish a common data schema that enables services to operate on or within any of these applications. This entails taking any of the data flowing through or about any of these entities 652 and pooling the data into a framework through which other applications across supply and demand can interact with the entities 652. can do. This may be a shared data pipeline originating from IoT systems and other external data sources, fed into a monitoring layer, and stored as a common data schema in a storage layer, where various intelligences identify impacts across these entities 652. can be trained to do so. In an example embodiment, a determination is made that the provider may be insolvent, or that the provider has become insolvent, at which point VCNP 604 will automatically trigger a replacement smart contract to be sent to the secondary provider with the changed terms. You can. There may be different aspects of management of the supply chain. For example, immediately and automatically change prices on the demand side in response to one or more suppliers being identified as bankrupt (e.g., from a bankruptcy announcement). Other similar examples may be used based on what happens in that automation layer that may be enabled by VCNP 604. Then, at the interface layer of this VCNP 604, the digital twin can view all of these entities 652, which are typically not shown together, and monitor what is going on in each of these entities 652, including identification of problem states. Users can use it to do this. For example, after seeing three quarters of poor financial reporting for a supplier, the report may be flagged to be watched closely for potential future bankruptcies, etc.

For example, an IoT system deployed at a fulfillment center 628 may be coordinated with an intelligent product 1510 that takes customer feedback regarding a product 1510, and an application 630 for the fulfillment center 628 may: Upon receiving customer feedback via the path to intelligent product 1510 regarding an issue with product 1510, a similar product 650 may be requested prior to product 650 being dispatched from fulfillment center 628. A workflow may be initiated to perform corrective action. Similarly, port infrastructure facilities 660, such as yards for holding shipping containers, provide connectivity to floating assets 620 (such as ships, barges, etc.) to accommodate the fleet of floating assets 620 to the port's capacity limits. may be notified that it is close, thereby initiating a negotiation process for remaining capacity (which may include automated negotiations based on a set of rules and governed by a smart contract) and allowing some assets 620 to move to alternative ports. or redirected to a holding facility. These and many other connections between value chain network entities 652 may be one-to-one connections, one-to-many connections, many-to-many connections, or defined connections of entities 652 (e.g., controlled by the same owner or operator). Any connections between groups are included in this application as applications 630 managed by VCNP 604.

VALUE CHAIN NETWORK ACTIVITIES AND APPLICATIONS MANAGED BY THE PLATFORM

Referring to FIG. 10 , a set of value chain network entities 652 are provided on VCNP 604, integrated with VCNP 604, and/or managed by or for VCNP 604. The set of applications 614 may include, without limitation, any one or more of the following broad types of applications: Supply chain management applications 812 (e.g., without limitation, product, component, and other To manage the timing, quantity, logistics, shipping, delivery, and other details of orders for items); Asset management application 21004 (e.g., without limitation, used for location of floating assets (such as ships, boats, barges, and floating platforms), warehouses, ports, shipyards, distribution centers, and other buildings) Real property, equipment, machinery and fixtures (such as containers, cargo, packages, goods, and other items used to handle them), (such as forklifts, delivery trucks, autonomous vehicles, and other systems used to move items) ) for managing value chain assets (such as vehicles, human resources (such as workers), software, information technology resources, data processing resources, data storage resources, power generation and/or storage resources, computational resources and other assets); Financial applications 822 (for handling financial matters related to value chain entities and assets, such as, but not limited to, involving payments, securities, collateral, receivables, duties, duties, import duties, taxes, etc.); Risk management applications 818 (e.g., without limitation, shipments, goods, products, assets, people, floating assets, vehicles, equipment items, components, information technology systems, security systems, security events, cybersecurity systems, property items, To manage risks or liabilities relating to health conditions, death, fire, flood, weather, disability, negligence, business interruption, personal injury, property damage, business damage, breach of contract, etc.); Demand management applications 824 (e.g., without limitation, demand planning applications, demand forecasting applications, sales applications, future demand aggregation applications, marketing applications, advertising applications, e-commerce applications, marketing analytics applications, customer relationship management applications, search engine optimization) Applications, sales management applications, advertising network applications, behavior tracking applications, marketing analytics applications, location-based product or service targeting applications, collaborative filtering applications, recommendation engines for products or services, etc., and one or more features of intelligent products 1510 Analyze interest by customers in categories of goods that can be supplied by the facility of value chain products or services, such as those used or enabled by or otherwise implemented using intelligent capabilities on intelligent products 1510. , planning, or facilitating applications); Trading application 858 (e.g., without limitation, a buying application, a selling application, a bidding application, an auction application, a reverse auction application, a bid/ask matching application, an analytics application to analyze value chain performance, yield, return on investment, or other metrics) etc); Tax applications 850 (e.g., without limitation, taxes, duties, import duties, levies, duties, duties, credits, fees or government imposed charges, such as, without limitation, customs duties, value-added taxes, sales taxes, income taxes, property taxes, municipalities) to manage, calculate, report, optimize or otherwise handle data, events, workflows or other factors related to fees, pollution taxes, renewable energy credits, pollution abatement credits, import duties, export duties, etc.); Identity management applications 830 (e.g., identity verification application, biometric identification verification application, pattern-based identity verification application, location-based identity verification application, user behavior-based application, fraud detection application, network address-based One of the entities 652 involved in the value chain, such as, but not limited to, one or more of a fraudulent transaction detection application, a black list application, a white list application, a content inspection-based fraudulent transaction detection application, or other fraudulent transaction detection applications. to manage the identity of the above); Inventory management application 820 (for example, without limitation, for managing inventory at a fulfillment center, distribution center, warehouse, storage facility, store, port, vessel or other floating asset, or other location); A security application, solution or service 834 (referred to herein as a security application, such as, but not limited to, any of the identity management applications 830 mentioned above, as well as physical security systems (e.g., access control systems (e.g. , biometric access control, fingerprinting, retina scanning, passwords, and other access controls), for safes, vaults, cages, security rooms, secure storage facilities, etc.), monitoring systems (e.g. cameras, motion sensors, using infrared sensors and other sensors), perimeter security systems, floating security systems for floating assets, cybersecurity systems (e.g. virus detection and remediation, intrusion detection and remediation, spam detection and remediation, phishing detection and remediation, social engineering detection and remediation, cyber attack detection and remediation, packet inspection, traffic inspection, DNS attack remediation and detection, etc.) or other security applications); Safety applications 840 (e.g., without limitation, to improve worker safety, to reduce the likelihood of damage to property, to reduce the risk of accidents, to reduce the likelihood of damage to goods (e.g., cargo) for risk management associated with insurance items, collateral for loans, etc., including security management based on any of the data sources, events or entities mentioned throughout this disclosure or documents incorporated by reference herein; including any application to detect, characterize or predict the likelihood and/or extent of an accident or other damaging event); Blockchain application 844 (e.g., a distributed ledger that captures a series of transactions, such as, without limitation, withdrawals or credits, buys or sells, exchanges for in-kind consideration, smart contract events, etc., or other blockchain-based applications); Facilities management applications 850 (e.g., infrastructure, buildings, and systems involved in supporting the value chain, such as, but not limited to, shipyards, ports, distribution centers, warehouses, docks, stores, fulfillment centers, storage facilities, etc.) In addition to managing real estate, personal property, and other property, it also includes information technology systems, robotic/autonomous vehicle systems, packaging systems, packing systems, picking systems, inventory tracking systems, inspection systems, and routing systems for mobile robots; for the design, management or control of systems and facilities within or around property, such as workflow systems for human assets, etc.); Regulatory applications 852 (e.g., without limitation, permitted routes, permitted freight and goods, permitted parties to transactions, mandatory disclosures, privacy, pricing, marketing, offerings of goods and services, use of data (data privacy regulations) Applications, services, transactions, activities, workflows referred to in this application and the documents incorporated by reference in this application, such as regulations of banking, marketing, sales, financial planning, etc., including regulations related to storage of data, etc. , an application for regulating any of the events, entities, or other items); commerce applications, solutions or services 854 (e.g., without limitation, an e-commerce site marketplace, online site, auction site or marketplace, physical product marketplace, advertising marketplace, reverse auction marketplace, advertising network, or other marketplace); Vendor management applications 832 (as long as they can be supplied in the value chain, such as, but not limited to, those involving features such as vendor qualification, vendor evaluation, request for proposal, request for information, receivables or other assurances of performance, contract management, etc. an application for managing the procurement of goods, components or materials of a set and/or for managing a set of vendors or prospective vendors); Analytics applications 838 (e.g., without limitation, big data applications, user behavior applications, prediction applications, classification applications, dashboards, pattern recognition applications, econometric applications, financial yield applications, return on investment applications, scenario planning applications, decision support Any of the data types, applications, events, workflows, or entities mentioned throughout this disclosure or documents incorporated by reference in this application, such as applications, demand forecasting applications, demand planning applications, route planning applications, weather forecasting applications, etc. Analysis application for things of); Pricing application 842 (e.g., for pricing of goods, services, including without limitation any mentioned throughout this disclosure and documents incorporated by reference herein); and smart contract applications, solutions, or services (collectively referred to in this application as smart contract applications 848, such as, without limitation, smart contracts for the sale of goods, smart contracts for ordering goods, smart contracts for delivery resources). Smart contracts, smart contracts for workers, smart contracts for the delivery of goods, smart contracts for the installation of goods, smart contracts using tokens or cryptocurrencies for consideration, rights, options, future or Smart contracts that grant interest, smart contracts for securities, commodities, futures, options, derivatives, etc., smart contracts for current or future resources, smart contracts structured to take into account or accept tax, regulatory or compliance parameters, arbitrage. any of the smart contract types referenced in this disclosure or in documents incorporated by reference in this application, such as a smart contract configured to execute a transaction). Accordingly, the value chain management platform 604 may host enabling interactions between a wide range of heterogeneous applications 630 (such term includes the above-referenced and other value chain applications, services, solutions, etc.), and thus , shared microservices, shared data infrastructure, and shared intelligence, any pair or larger combination or permutation of these services can be improved over isolated applications of the same type.

Still referring to FIG. 10 , it involves a set of value chain network entities 652 that are provided on VCNP 604, integrated with VCNP 604, and/or managed by or for VCNP 604. The set of applications 614 may further include, without limitation: a payment application 860 (e.g., for any of the applications 630 mentioned herein) (geography of entity 652 (including based on situational factors such as applicable taxes, duties, etc.) for calculating payments, transferring funds, settling payments to parties, etc.); Process management application 862 (e.g., of the processes or workflows described throughout this disclosure, including supply process, demand process, logistics process, delivery process, order fulfillment process, distribution process, order process, navigation process, etc. to manage random things); For example, a compatibility test application 864 to assess compatibility between value chain network entities 652 or activities involved in any of the processes, workflows, activities, or other applications 630 described herein (e.g. , compatibility of the container or package with the product 1510, compatibility of the product 1510 with a set of customer requirements, compatibility of the product 1510 with another product 1510 (e.g., refill, refill of one for the other). supply, replacement parts, etc.), compatibility of infrastructure and equipment entities 652 (e.g., between container ships or barges and ports or waterways, between containers and storage facilities, between trucks and roads, between drones or robots and packages, drones, between AV or robot and delivery destination, etc.); Infrastructure test applications 802 (e.g., the ability of an infrastructure element to support product 1510 or application 630 (e.g., without limitation, storage capacity, lifting capacity, movement capacity, storage capacity, network capacity, environment) to test control capabilities, software capabilities, security capabilities, etc.); and/or an incident management application 910 (e.g., vehicle accidents, worker injuries, shutdown incidents, property damage incidents, product damage incidents, product liability incidents, regulatory non-compliance incidents, health and/or safety incidents, traffic congestion and/or delays). Incidents (including network traffic, data traffic, vehicle traffic, maritime traffic, human operator traffic, etc., as well as combinations thereof), product failure incidents, system failure incidents, system performance incidents, fraud incidents, misuse incidents, unauthorized To manage events, incidents, and other incidents that may occur in one or more environments involving value chain network entities 652, such as, but not limited to, usage incidents, etc.).

Still referring to FIG. 10 , a set of value chain network entities 652 is provided on VCNP 604, integrated with VCNP 604, and/or managed by or for VCNP 604. The set of applications 614 may further include, without limitation: predictive maintenance applications 910 (e.g., value chain network entities 652 such as products 650, equipment, infrastructure, buildings, vehicles, etc. To anticipate, anticipate, and take actions to manage defects, failures, shutdowns, damage, required maintenance, required repairs, required services, required support, etc., for a set of ); Logistics applications 912 (e.g., loading, transfer of goods to pickup, delivery, and transport facilities involved in scheduling and managing the movement of products 650 and other items between points of origin and destination through various intermediate locations) to manage logistics for unloading, packing, picking, shipping, driving, and other activities); Reverse logistics applications 914 (e.g., for handling logistics for returned products 650, waste, damaged goods, or other items that may be transported on the return logistics path); Waste reduction applications 920 (e.g., packaging waste, solid waste, energy waste, liquid waste, contamination, contaminants, waste of computing resources, waste of human resources, or other waste reduction applications involving value chain network entities 652 or activities) to reduce waste); Augmented reality, mixed reality and/or virtual reality applications 930 (e.g., in the movement of product 1510, the interior of a facility, the state or condition of an item of goods, one or more environmental conditions, weather conditions, a container or set of containers) for visualizing activities involved in one or more of one or more value chain network entities 652 or applications 630, such as, but not limited to, packing configuration for, etc.); Demand forecasting application 940 (e.g., predicts demand for product 1510, categories of products, potential products, and/or factors accompanying demand such as market factors, wealth factors, demographic factors, weather factors, economic factors, etc. to predict); Demand aggregation application 942 (e.g., information about one or more products 650, categories, etc., including current demand for existing products and future demand for products not yet available, orders and/or commitments (smart contracts) to aggregate (optionally implemented in one or more contracts); Customer profiling applications 944 (e.g., past purchase data, loyalty program data, behavior tracking data (including data captured from the customer's interaction with smart product 1510), online clickstream data, intelligent agents to profile one or more demographic, psychographic, behavioral, economic, geographic, or other attributes of a set of customers, including based on interactions with, and other data sources); and/or a component supply application 948 (e.g., for managing the supply chain of components for a set of products 650).

Still referring to FIG. 10 , it involves a set of value chain network entities 652 that are provided on VCNP 604, integrated with VCNP 604, and/or managed by or for VCNP 604. The set of applications 614 may further include, without limitation: a policy management application 868 (e.g., one or more policies for governance of one or more value chain network entities 652 or applications 630); , to deploy rules, etc.), e.g. to govern the execution of one or more workflows (which may involve configuring policies in platform 604 on a workflow-specific basis), regulatory (ocean, food and To govern compliance with drug, medical, environmental, health, safety, tax, financial reporting, commercial, and other regulations as described throughout this disclosure or as understood in the art), resources (e.g. , to govern the provisioning of (connectivity, computing, human, energy, and other resources), to govern compliance with company policies, and to govern contracts (e.g., platform 604, via connectivity facility 642, To govern compliance of entities 652 and applications 630 (including smart contracts that can automatically deploy governance features), and to govern interactions with other entities (e.g., sharing of information). and policies for access to resources), to govern data access (including privacy data, operational data, state data, and many other data types), infrastructure, products, equipment, locations, etc. (to govern secure access to); Product configuration application 870 (e.g., to allow a product manager and/or an automated product configuration process (optionally using robotic process automation)) to determine configuration for product 1510, e.g., in an agile manner. Involves on-the-fly configuration during manufacturing, or involves configuration or customization in route (e.g., by 3D printing one or more features or elements), or involves, for example, downloading firmware. (involving remote configuration or customization by configuring field programmable gate arrays, installing software, etc.); Warehousing and fulfillment applications 872 (e.g., selecting products, configuring storage locations for products, determining routes for personnel, mobile robots, etc. to move products around the facility, picking, and Determining packing schedules, routes, and workflows; managing the operation of robots, drones, conveyors, and other facilities; determining schedules for moving products to shipping docks; and many other functions. (for managing warehouses, distribution centers, fulfillment centers, etc., such as those involving); Kit configuration and deployment applications 874 (e.g., allow a user, customer, or operator of the VCNP to rapidly deploy a subset of the capabilities of the VCNP 604 for a specific value chain network entity 652 and/or application 630 to enable configuring a pre-integrated, pre-provisioned, and/or pre-configured system in a kit, box, or other manner to allow for: and/or a product test application 878 to test the product 1510 (for performance, capabilities and enablement of features, safety, compliance with policies or regulations, quality, quality of service, probability of failure, and many other factors). included).

Still referring to FIG. 10 , a set of value chain network entities 652 is provided on VCNP 604, integrated with VCNP 604, and/or managed by or for VCNP 604. The set of applications 614 may further include, without limitation: a maritime fleet management application 880 (e.g., determining optimal routes for fleet assets based on weather, markets, traffic, and other conditions) A set of maritime assets such as container ships, barges, boats, etc., as well as docks, cranes, etc., to ensure compliance with policies and regulations, ensure safety, improve environmental factors, improve financial metrics, etc. , to manage related infrastructure facilities such as ports, etc.); Delivery management applications 882 (e.g., to optimize financial yield, improve safety, reduce energy consumption, reduce delays, mitigate environmental impacts, and for many other purposes, e.g., trucks, trains, airplanes, etc.) to manage a set of shipping assets, etc.); Opportunity matching application 884 (e.g., for matching one or more demand factors with one or more supply factors, for matching the needs and capabilities of value chain network entities 652, for identifying reverse logistics opportunities, to identify opportunities for inputs to enrich analytics, artificial intelligence and/or automation, to identify cost reduction opportunities, to identify profit and/or arbitrage opportunities, etc.); Workforce management applications 888 (e.g., fulfillment centers, ships, ports, warehouses, distribution centers, corporate management locations, retail stores, online/e-commerce site management facilities, ports, ships, boats, barges, trains, warehouses, and to manage a diverse workforce of workers, including the workforce at, on, or for other facilities referenced throughout this disclosure); Distribution and delivery applications 890 (e.g., for planning, scheduling, routing, and otherwise managing the distribution and delivery of products 650 and other items); and/or enterprise resource planning (ERP) applications 892 (e.g., for planning the utilization of enterprise resources, including workforce resources, financial resources, energy resources, physical assets, digital assets, and other resources).

CORE CAPABILITIES AND INTERACTIONS OF THE DATA HANDLING LAYERS (ADAPTIVE INTELLIGENCE, MONITORING, DATA STORAGE AND APPLICATIONS)

11, a high-level schematic diagram of an embodiment of a value chain network management platform 604 is illustrated, which may occur within one or more value chain network processes, workflows, activities, events and/or applications 630. , a system, application, or process that operates in concert to enable intelligent management of a set of value chain entities 652 that are capable of operating, transacting, etc., or that operate, support, or enable them on their own. , may include a set of modules, services, layers, devices, components, machines, products, subsystems, interfaces, connections, and other elements, or may otherwise be used in product 1510 (in embodiments, processing, networking, sensing, computing, , and/or other IoT-enabled intelligent products, which may include finished products, software products, hardware products, component products, materials, items of equipment, consumer packaging products, consumer products, food products, beverage products, is part of or integrated with Platform 604 (which may be a household product, business supply product, consumable product, pharmaceutical product, medical device product, technology product, entertainment product, or any other type of product or related service) It may be, be associated with, or be operated by. A wide range of value chain activities (e.g., supply chain activities, logistics activities, demand management and planning activities, delivery activities, delivery activities, warehouse activities, etc.), as involved in various value chain network processes, workflows, activities, events and applications 630. Value chain entities 652, such as those accompanying or for house activities, distribution and fulfillment activities, inventory counting, storage and management activities, marketing activities, etc., are described throughout this disclosure or in documents incorporated by reference herein. It may include any of a wide variety of assets, systems, devices, machines, components, equipment, facilities, individuals or other entities mentioned.

In embodiments, the value chain network management platform 604 may include a set of data handling layers 608, each of which may be used for a wide variety of value chain network applications and end uses, such as automation, machine learning, artificial intelligence, etc. Configured to provide a set of capabilities that facilitate the development and deployment of intelligence, to facilitate the application of intelligence, intelligent transactions, intelligent actions, remote control, analysis, monitoring, reporting, state management, event management, process management, etc. do. In an embodiment, the data handling layer 608 includes a value chain network monitoring system layer 614, a value chain network entity-oriented data storage system layer 624 (referred to simply as data storage layer 624 in some cases in this application for convenience). ), an adaptive intelligence system layer 614, and a value chain network management platform 604. Value chain network management platform 604 may include a data handling layer 608, such that value chain network management platform 604 may be a management and/or value chain network monitoring system layer of value chain network management platform 604. 614, the value chain network entity-oriented data storage system layer 624 (e.g., data storage layer 624), and the adaptive intelligence system layer 614. Each of the data handling layers 608 may include various services, programs, applications, workflows, systems, components and modules, as further described in this application and documents incorporated by reference herein. In embodiments, each of the data handling layers 608 (and optionally the entire platform 604) is configured such that one or more of its elements can be accessed as a service by other layers 624 or by other systems (e.g. (e.g., configured as a platform-as-a-service deployed on a set of cloud infrastructure components in a microservices architecture). For example, platform 604 may include network connections (including various configurations, types, and protocols), interfaces, ports, application programming interfaces (APIs), brokers, services, connectors, wired or wireless communication links, and human-accessible may have (or may configure and/or provision) a set of connectivity facilities 642 such as interfaces, software interfaces, micro-services, SaaS interfaces, PaaS interfaces, IaaS interfaces, cloud capabilities, etc., and a data handling layer 608 ) may use this, whereby data or information is transferred between the data handling layer 608 and other layers, systems or subsystems of the platform 604, as well as to other systems or clouds, such as value chain entities 652. Can be exchanged with external systems, such as on-premises or on-premise enterprise systems (e.g., accounting systems, resource management systems, CRM systems, supply chain management systems, etc.). Each data handling layer 608 may include a set of services (e.g., microservices) for data handling, including facilities for data extraction, transformation, and loading; data cleaning and deduplication facilities; data normalization facility; data synchronization facility; data security facilities; Computational facilities (e.g., for performing predefined computational operations on data streams and providing output streams); compression and decompression facilities; analytics facilities (e.g., providing automated creation of data visualizations), etc.

In an embodiment, each data handling layer 608 has a set of application programming connection facilities 642 to automate data exchange with each of the other data handling layers 608. They represent data as it is exchanged between layers and/or applications 630, such as converting data from one format or protocol to another as needed for one layer to consume output from another layer. It may include data integration capabilities to, for example, extract, transform, load, normalize, compress, decompress, encode, decode, and otherwise process packets, signals, and other information. In an embodiment, the data handling layer 608 is structured in a topology that facilitates the collection and distribution of shared data across multiple applications and uses within the platform 604 by the value chain monitoring system layer 614. Value chain monitoring system layer 614 is for collecting and organizing data collected from or about value chain entities 652, as well as data collected from or about various data layers 624 or services or components. It may include, integrate with, and/or cooperate with various data collection and management systems 640, referred to in some cases for convenience as data collection system 640. For example, a stream of physiological data from a wearable device worn by a worker performing a task or a consumer engaging in an activity can be transmitted through the monitoring system layer 614 to a number of distinct applications within the value chain management platform 604; For example, applications that facilitate monitoring a worker's physiological, psychological, performance level, alertness, or other conditions and other applications that facilitate operational efficiency and/or effectiveness. In embodiments, monitoring system layer 614 facilitates alignment, such as time synchronization, normalization, etc., of data collected for one or more value chain network entities 652. One or more video streams or other sensor data collected by or about workers 718 or other entities within a value chain network facility or environment, for example, from a set of camera-enabled IoT devices, are aligned with a common clock. The relative timing of a set of video or other data can be understood by a system capable of processing video, such as a machine learning system that operates on images within a video, changes between images within different frames of a video, etc. . In this example, monitoring system layer 614 monitors a stream of data from wearable devices (ships, lifts, vehicles, containers, cargo handling systems, packing systems, delivery systems, drones, etc.). It can be further aligned with other data, such as streams of data generated by value chain network systems (such as fields/robots, etc.), streams of data collected by mobile data collectors, etc. The organization of the monitoring system layer 614 as a set of common platforms or microservices accessed across many applications can be configured to support the value chain to have a growing set of IoT devices and a growing set of applications that monitor other systems and devices under their control. It can dramatically reduce the number of interconnections required by owners or other operators within the network.

In an embodiment, the data handling layer 608 is a platform 604 by value chain network-oriented data storage system layer 624, which in some cases is referred to herein for convenience simply as the data storage layer 624 or storage layer 624. ) consists of a topology that facilitates sharing or common data storage across multiple applications and uses. For example, various data collected about value chain entities 652 as well as data generated by other data handling layers 608 may be stored in data storage layer 624 so that various data handling layers 608 Any of the services, applications, programs, etc. may access a common data source (which may include a single logical data source distributed across disparate physical and/or virtual storage locations). This facilitates a dramatic reduction in the amount of data storage required to handle the enormous amounts of data generated by or about value chain network entities 652 as applications 630 and use of the value chain network grow and proliferate. can do. For example, a supply chain or inventory management application in the value chain management platform 604, such as for ordering replacement parts for an item of machinery or equipment, may determine the likelihood that a ship's component or a port's facility will require a replacement part. Predictive maintenance applications can access the same data set about which parts have been replaced for a set of machines. Similarly, forecasts may be used in connection with resupply of items.

In embodiments, value chain network data objects 1004 are classes, objects, properties, etc. of the set of data objects (such as those associated with value chain network entities 652 and applications 630) handled by platform 604. It may be provided according to an object-oriented data model that defines parameters and other features.

In embodiments, data storage system layer 624 may be an expert system, an analytics system, an artificial intelligence system, a robotic process automation system, a machine learning system, a deep learning system, a supervised learning system, or any of the systems incorporated by reference in this disclosure and application. It can provide an extremely rich environment for the collection of data that can be used for extraction of features or inputs to an intelligent system, such as other intelligent systems as disclosed throughout the document. As a result, each application 630 within platform 604 and each adaptive intelligence system within adaptive intelligence system layer 614 can benefit from data collected or generated by or for each of the others. . In embodiments, data storage system layer 624 may facilitate the collection of data that can be used for extraction of features or inputs to an intelligent system, such as a development framework from artificial intelligence. In an example, the collection of data may include pulling in and/or accepting event logs (naturally stored or ad-hoc, as needed), or periodically accessing onboard diagnostic data. You can perform inspections, etc. In examples, pre-computation of features may be deployed using, for example, AWS Lambda, or various other cloud-based on-demand computing capabilities, such as pre-computing, multiplexing signals. In many instances, even when collecting many, many data types from relatively heterogeneous entities, one or more sets of capabilities of the data handling layer 608 can be used to provide connectivity and connectivity across value chain entities and across applications used by the entities. There are similar types of pairings of value chain entities (double, triple, quadruple, etc.) on which services can be deployed. In these examples, various pairings of similar types of value chain entities that at least partially utilize connectivity and services across value chain entities and applications can utilize information from the pairings of connected data to utilize the various neural networks disclosed in this application and hybrid combinations thereof. It can be directed to artificial intelligence services, including: In this example, genetic programming techniques may be deployed to prune some of the input features within the information from pairings of linked data. In these examples, genetic programming techniques may also be deployed to add and augment input features from information from the pairing. These genetic programming techniques could be shown to increase the efficacy of decisions made by artificial intelligence services. In this example, information from the pairing of linked data can be used to facilitate the shared data schema as a capability and resource for platform 604, including supporting or deploying robotic process automation, forecasting, forecasting, and other resources. It can be migrated to other layers on the platform.

A wide range of data types can be stored in the storage layer 624 using a variety of storage media and data storage types, data architectures 1002, and formats, which can be used to store many of the data associated with the value chain network entities 652 and applications 630. Among other data types include, without limitation: asset and facility data 1030, state data 1140 (e.g., any of the value chain network entities 652, applications 630 or components or workflows, among others) (indicating a state, condition state, or other indicator for any of the components or elements of platform 604), worker data 1032 (identity data, role data, task data, workflow data) , health data, attention data, mood data, stress data, physiological data, performance data, quality data and many other types); Event data 1034 (e.g., for any of a wide range of events, including operational data, transaction data, workflow data, maintenance data, and events occurring within the value chain network 668, or many related thereto) Containing different types of data or for one or more applications 630, process events, financial events, transaction events, output events, input events, state change events, operational events, workflow events, repair events, maintenance events, including service events, damage events, injury events, replacement events, refueling events, recharge events, delivery events, warehousing events, transfers of goods, border crossings, movement of cargo, inspection events, supply events, etc.); Claims data 664 (relating to insurance claims, such as for business interruption insurance, product liability insurance, insurance for goods, facilities or equipment, flood insurance, insurance for contract-related risks, etc., as well as product liability, general Claims data relating to liability, workers' compensation, personal injury and other liability claims, and supply contract performance claims, product delivery requirements, warranty claims, indemnity claims, delivery requirements, timing requirements, milestones, key performance indicators, etc. billing data relating to the same contract); accounting data 730 (e.g., data regarding completion of contract requirements, satisfaction of receivables, payment of duties and customs, etc.); and risk management data 732 (e.g., regarding items supplied, quantity, price, delivery, source, route, tariff information, etc.).

In embodiments, data handling layer 608 is configured in a topology that facilitates shared adaptive capabilities, which in some cases are referred to in this application as adaptive intelligence layer 614 for convenience. It may be provided, managed, mediated, etc. by one or more of a set of services, components, programs, systems, or capabilities in layer 614. Adaptive intelligence systems layer 614 may include a set of data processing, artificial intelligence, and computational systems 634 described in greater detail elsewhere throughout this disclosure. Accordingly, computing resources (such as available processing cores, available servers, available edge computing resources, available on-device resources (for a single device or peered network), and available cloud infrastructure, among others), ( Local storage on the device, storage resources within or on a value chain entity or environment (including on-device storage, storage on asset tags, local area network storage, etc.), network storage resources, cloud-based storage resources, database resources, etc. (including data storage resources), networking resources (including cellular network spectrum, wireless network resources, fixed network resources, etc.), energy resources (such as available battery power, available renewable energy, fuel, grid-based power, etc.), etc. The use of the same various resources may be optimized in a coordinated or shared manner on behalf of an operator, enterprise, etc., for the benefit of, for example, multiple applications, programs, workflows, etc. For example, the adaptive intelligence layer 614 could manage and provision available network resources for both supply chain management applications and demand planning applications (among many other possibilities), so that low-latency resources are distributed to supply chain management. It is used for applications (where quick decisions can be important), while longer latency resources are used for demand planning applications. As described in greater detail throughout this disclosure and the documents incorporated by reference herein, various services and capabilities across various tiers 624 may be provided, including application requirements, quality of service, timely delivery, service objectives, budgets; A wide variety of adaptations can be provided, including those based on cost, price, risk factors, operational objectives, efficiency objectives, optimization parameters, return on investment, profitability, uptime/downtime, operator utilization, etc.

Value chain management platform 604, referred to in some instances herein as platform 604 for convenience, allows operators to store common data in a data storage layer 624, collect or monitor common data in a monitoring system layer 614, and /or value in a common application environment (e.g., shared, pooled, similar licenses - whether shared data for one person, multiple people, or anonymous), such as leveraging the common adaptive intelligence of the adaptive intelligence layer 614 Includes, and is integrated with, various value chain network processes, workflows, activities, events and applications 630 described throughout this disclosure that enable management of two or more aspects of a chain network environment or entity 652 , can make these possible. Output from applications 630 within platform 604 may be provided to other data handling layers 624. These include, without limitation, status and status information for various objects, entities, processes, flows, etc.; Object information such as identity, properties, and parameter information for various classes of objects of various data types; Event and change information for other flows, including, for example, workflows, dynamic systems, processes, procedures, protocols, algorithms, and timing information; In particular, indications of success and failure, indications of completion of a process or milestone, indications of accurate or inaccurate forecasts, indications of accurate or inaccurate labeling or classification, and success metrics (yield, engagement, return on investment, profitability, efficiency, timeliness, It may include outcome information such as service quality, product quality, customer satisfaction, etc.). The output from each application 630 may be stored in the data storage layer 624, distributed for processing by the data collection layer 614, and used by the adaptive intelligence layer 614. Accordingly, the cross-application nature of platform 604 provides, for example, machine learning for results across applications and automation of a given application through machine learning based on results from other applications or other elements of platform 604. All necessary infrastructure for adding intelligence to any given application, providing enhancements and allowing application developers to focus on application-native processes while benefiting from other capabilities of the platform 604. Facilitates convenient organization of elements. In examples, systems, components, services and other capabilities that control, automate, or optimize one or more performance characteristics of one or more value chain network entities 652; Or, capabilities may exist to generally improve any of the processes and application outputs and results 1040 sought by use of the platform 604. In some examples, output and results 1040 from various applications 630 may be used to facilitate automated learning and improvement, such as classification, prediction, etc., involved in the steps of the process that are intended to be automated.

SOME DATA STORAGE LAYER DETAILS - ALTERNATIVE DATA ARCHITECTURES

12, additional details, components, subsystems, and other elements of an optional embodiment of the data storage layer 624 of platform 604 are illustrated. Traditional relational and object-oriented data architectures, blockchain architecture (1180), asset tag data storage architecture (1178), local storage architecture (1190), network storage architecture (1174), multi-tenant architecture (1132), distributed data Architecture (1002), Value Chain Network (VCN) Data Object Architecture (1004), Cluster-Based Architecture (1128), Event Data-Based Architecture (1034), State Data-Based Architecture (1140), Graph Database Architecture (1124) A variety of data architectures may be used, including , self-organizing architecture 1134, and other data architectures 1002.

The adaptive intelligence system layer 614 of the platform 604 can support a variety of different data storage architectures 1002, such as allowing extraction from one type of database and loading into a data system using a different protocol or data structure. ) may include one or more protocol adapters 1110 to facilitate data storage, search access, query management, loading, extraction, normalization, and/or transformation.

In an embodiment, the value chain network-oriented data storage system layer 624 may include physical storage systems, virtual storage systems, local storage systems (e.g., part of local storage architecture 1190), distributed storage systems, databases, memory, may include, without limitation, network-based storage, network attached storage systems (e.g., part of network storage architecture 1174 such as those using NVME, storage attached networks, and other network storage systems), and the like.

In an embodiment, storage layer 624 may be one or more knowledge graphs (e.g., directed acyclic graphs, data maps, data hierarchies, data clusters containing links and nodes, self-organizing maps, etc.) within graph database architecture 1124. Data can be stored in . In an example embodiment, a knowledge graph may be a common example of when graph databases and graph database architectures can be used. In some examples, a knowledge graph can be used to graph workflows. For linear workflows, a directed acyclic graph can be used. For conditional workflows, a cyclic graph can be used. A graph database (e.g., graph database architecture 1124) may include a knowledge graph, or a knowledge graph may be an example of a graph database. In an example embodiment, a knowledge graph may include an ontology and connections (e.g., relationships) between ontologies of the knowledge graph. In an example, a knowledge graph could be used to capture an articulation of a human expert's knowledge domain to identify opportunities to design and build robotic process automation or other intelligence that can replicate this knowledge set. The platform can be used to recognize what kind of expert uses this factual knowledge base (from the knowledge graph) combined with capabilities that can be replicated by artificial intelligence, which can differ depending on the type of expertise involved. . For example, artificial intelligence, such as convolutional neural networks, can be used with spatiotemporal aspects that can be used to diagnose problems or pack boxes in a warehouse. On the other hand, the platform may use a different type of knowledge graph for self-organizing maps of experts whose main task is to segment customers into customer segmentation groups. In some examples, a knowledge graph can be built from a variety of data, such as job credentials, job listings, and parse output. In embodiments, data storage layer 624 may store data in digital threads, ledgers, etc., for example, to maintain serial or other records of entities 652 over time, including any of the entities described in this application. can be saved. In embodiments, the data storage layer 624 may use and enable asset tags 1178, which may contain data structures associated with assets and accessible and managed, e.g., through the use of access controls; Accordingly, storage and retrieval of data is optionally associated with local processes, but is also optionally open to remote retrieval and storage options. In an embodiment, the storage layer 624 may be configured with access controls, which may be role-based or based on credentials associated with the value chain entity 652, service, or one or more applications 630, e.g. It may include one or more blockchains 1180 that store identity data, transaction data, past interaction data, etc. Data stored by data storage system 624 may include accounting and other financial data 730, access data 734, asset and facility data 1030 (e.g., for any of the value chain assets and facilities described herein). for), asset tag data 1178, operator data 1032, event data 1034, risk management data 732, pricing data 738, safety data 664, and incorporated herein by reference. may include many different types of data that may be associated with, generated by, or about any of the value chain entities and activities described in the document.

ADAPTIVE INTELLIGENT SYSTEMS AND MONITORING LAYERS

13, additional details, components, subsystems, and other elements of an optional embodiment of platform 604 are illustrated. The management platform 604 may, in various optional embodiments, include a set of applications 614 that enable operators or owners of value chain network entities, or other users, to access the relevant applications 614 above and throughout this disclosure. may manage, monitor, control, analyze, or otherwise interact with one or more elements of value chain network entity 652, such as any of the elements mentioned.

In embodiments, adaptive intelligence system layer 614 is a set of systems, components, services and other capabilities that collectively facilitate the coordinated development and deployment of intelligent systems, e.g., applications 630 in application platform 604. that can improve one or more of the following; capable of improving the performance of one or more components, or the overall performance of the connectivity facility 642 (e.g., speed/latency, reliability, quality of service, cost reduction, or other factors); capable of improving other capabilities within the adaptive intelligence system layer 614; improving the performance (e.g., speed/latency, energy utilization, storage capacity, storage efficiency, reliability, security, etc.) or overall performance of one or more components of the value chain network-oriented data storage system 624; controlling, automating, or optimizing one or more performance characteristics of one or more value chain network entities 652; or generally improving any of the processes and application outputs and results 1040 sought by use of the platform 604.

This adaptive intelligence system 614 includes a robotic process automation system 1442, a set of protocol adapters 1110, a packet acceleration system 1410, an edge intelligence system 1420 (which may be a self-adaptive system), an adaptive A networking system 1430, a set of state and event managers 1450, a set of opportunity miners 1460, a set of artificial intelligence systems 1160, a set of digital twin systems 1700, an entity interaction system. (1920) (e.g., for setting up, provisioning, configuring and otherwise managing a set of interactions between a set of value chain network entities 652 within a value chain network 668), and other systems. It can be included.

In embodiments, value chain monitoring system layer 614 and data collection system 640 may include a wide range of systems for collection of data. This layer may include, without limitation, real-time monitoring systems 1520 (e.g., vessels and other floating assets, delivery vehicles, trucks and other transport assets, and event and status reporting systems at shipyards, ports, warehouses, distribution centers and other locations). such as on-board monitoring system); On-board diagnostics (OBD) and telematics systems for floating assets, vehicles and equipment; A system that provides diagnostic codes and events via an event bus, communications port, or other communications system; monitoring infrastructure (e.g., cameras, motion sensors, beacons, RFID systems, smart lighting systems, asset tracking systems, people tracking systems, and ambient sensing systems located in various environments where value chain activities and other events occur), as well as Additionally, removable and replaceable monitoring systems such as portable and mobile data collectors, RFID and other tag readers, smartphones, tablets, and other mobile devices capable of data collection; Software interaction observation system 1500 (in particular, software interactions that occur as a result of another program, e.g., via an API, as well as mouse movements, touchpad interactions, mouse clicks, cursor movements, keyboard interactions) , for logging and tracking events accompanying a user's interaction with a software user interface, such as navigation actions, eye movements, finger movements, gestures, menu selections, etc.); mobile data collector 1170 (e.g., broadly described herein and in documents incorporated by reference), visual monitoring system 1930 (e.g., items, people, materials, components, machines, equipment, personnel, gestures, Video and still imaging systems, LIDAR, IR and other systems that allow visualization of the representation, location, position, configuration, and other factors or parameters of an entity 652, as well as inspection to monitor processes, operator activities, etc. using the system); Interaction point systems 1530 (e.g., dashboards, user interfaces, and control systems for value chain entities); Physical process observation system 1510 (e.g., physical activities of operators, workers, customers, etc., individuals (e.g., shippers, delivery workers, packers, pickers, assembly personnel, customers, merchants, vendors, distributors, and others)) For tracking physical activity, the physical interaction of workers with other workers, the interaction of workers with physical entities such as machines and equipment, and the interaction of physical entities with other physical entities, including, without limitation, video and still image cameras; Including through the use of motion detection systems (e.g., including optical sensors, LIDAR, IR and other sensor sets), robotic motion tracking systems (e.g., tracking the movement of systems attached to humans or physical entities), etc. ; Machine health monitoring system 1940 (any value chain entity, e.g., machine or component, e.g., machine, e.g., client, server, cloud resource, control system, display screen, sensor, camera, vehicle, robot, or other machine) (including on-board monitors and external monitors of conditions, states, operating parameters, or other measures of conditions); Sensors and Cameras 1950 and other IoT data collection systems 1172, including onboard sensors, value chain environments (e.g., without limitation, points of origin, loading or unloading docks, vehicles or floating assets used to transport goods, containers, Sensors or other data collectors (including click tracking sensors) in or around ports, distribution centers, storage facilities, warehouses, delivery vehicles, and destination points), cameras to monitor the overall environment, and certain Dedicated cameras for machines, processes, workers, etc., wearable cameras, portable cameras, cameras placed on mobile robots, cameras on portable devices such as smartphones and tablets, and cameras incorporated by reference throughout this disclosure or in this application. including many others including any of the many sensor types disclosed in the document); Indoor Location Monitoring System 1532 (cameras, IR systems, motion detection systems, beacons, RFID readers, smart lighting systems, triangulation systems, RF and other spectral detection systems, time-of-flight systems, chemical noses, and other chemical sensor sets) but also includes other sensors); User feedback system 1534 (including survey system, touch pad, voice-based feedback system, evaluation system, facial expression monitoring system, affect monitoring system, gesture monitoring system, etc.); Behavioral monitoring system 1538 (e.g., movement, shopping behavior, purchasing behavior, clicking behavior, behavior indicative of fraud or deception, user interface interaction, product return behavior, behavior indicative of interest, attention, boredom, etc., mood indications) to monitor behavior (e.g., fidgeting, staying still, moving closer, changing posture, etc.); and a wide variety of Internet of Things (IoT) data collectors 1172 such as those described throughout this disclosure and in documents incorporated by reference herein.

In an embodiment, the value chain monitoring system layer 614 and data collection system 640 include an entity discovery system for discovering one or more value chain network entities 652, such as any of the entities described throughout this disclosure. 1900). This can be done by device identifier, by network location, by geolocation (e.g. by geofencing), by indoor location (e.g. by proximity to known resources such as IoT-enabled devices and infrastructure, Wifi routers, switches, etc.). by), by cellular location (e.g., by proximity to a cellular tower), by an identity management system (e.g., by an identifier assigned and/or managed by platform 604), such as an owner, operator, , a user, or a company, if associated with another entity 652 ), and the like. Entity discovery 1900 may initiate a handshake between a set of devices, such as initiating interactions serving various applications 630 or other capabilities of platform 604.

14, a management platform of an information technology system, such as a management platform for a value chain of goods and/or services, is shown as a block diagram of functional elements and representative interconnections. The management platform includes, among other things, a user interface 3020 that provides a set of adaptive intelligence systems 614. Adaptive intelligence system 614 provides coordinated intelligence (artificial intelligence) for a set of demand management applications 824 and a set of supply chain applications 812 for categories of goods 3010 that can be produced and sold through the value chain. Provides intelligence (1160), expert systems (3002), machine learning (3004), etc.) Adaptive intelligence system 614 may deliver artificial intelligence 1160 through a set of data processing, artificial intelligence and computational systems 634. In embodiments, adaptive intelligence system 614 may operate on or within a set of value chain applications (e.g., demand management application 824 and supply chain application 812). It is selectable and/or configurable through user interface 3020 to operate in cooperation with. Adaptive intelligence system 614 may include any of a variety of expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, and other systems described throughout this disclosure and in documents incorporated by reference. It may include artificial intelligence, including.

In embodiments, the user interface may be configured to interact with selected members of the value chain to enhance, control, improve, optimize, configure, adapt, or otherwise influence the value chain for a category of goods 3010. Takes input from data sources (e.g., data sources used by set of demand management applications 824 and/or set of supply chain applications 812) and processes them, such as neural networks, artificial intelligence systems 1160 or throughout this disclosure. and may include an interface for configuring the artificial intelligence system 1160 to feed any of the other adaptive intelligence systems 614 described in the documents incorporated by reference herein. In embodiments, selected data sources in the value chain may be applied as input for classification or prediction, or as results related to the value chain, categories of products 3010, etc.

In embodiments, providing coordinated intelligence may include providing artificial intelligence capabilities, such as artificial intelligence system 1160 and the like. Artificial intelligence systems include, for example, value chain processes, material invoices, manifests, delivery schedules, weather data, traffic data, product design specifications, customer complaint logs, customer reviews, enterprise resource planning (ERP) systems, and customer relationship management (CRM). ) systems, Customer Experience Management (CEM) systems, Service Lifecycle Management (SLM) systems, Product Lifecycle Management (PLM) systems, etc., by processing data available from any of the data sources in the value chain, e.g. For a category, it may facilitate coordinated intelligence for a set of demand management applications 824 or a set of supply chain applications 812 or both.

In embodiments, user interface 3020 may provide access to artificial intelligence capabilities, applications, systems, etc., specifically for coordinating intelligence for value chain applications and specifically for value chain applications for categories of goods 3010. You can. User interface 3020 may be adapted to receive information describing the category of product 3010 and, in response thereto, configure user access to artificial intelligence capabilities such that the user, through the user interface, may receive information describing the category of product 3010. Guided by artificial intelligence capabilities suitable for use with value chain applications (e.g., a set of demand management applications 824 and supply chain applications 812) that contribute to the products/services within the category. User interface 3020 may facilitate providing coordinated intelligence, including artificial intelligence capabilities that provide coordinated intelligence for specific operators and/or enterprises participating in the supply chain for a category of goods.

In embodiments, user interface 3020 may be configured to facilitate a user selecting and/or configuring multiple artificial intelligence systems 1160 for use with a value chain. The user interface may present a set of demand management applications 824 and supply chain applications 812 as connected entities, each receiving, processing, and generating output that can be shared between the applications. A type of artificial intelligence system 1160 may be responsive to a set of connected applications or data elements being displayed in the user interface, such as by the user positioning the pointer proximate to the set of connected applications, etc. It can be. In embodiments, user interface 3020 may include supply chain application automation, demand management application automation, machine learning, artificial intelligence, intelligent transactions, intelligent behavior, remote control, analytics, monitoring, reporting, condition management, event management, and process management. may facilitate access to a set of adaptive intelligence systems that provide a set of capabilities that facilitate the development and deployment of intelligence for at least one function selected from a list of functions consisting of:

Adaptive intelligence system 614 may perform coordinated operations, e.g., when artificial intelligence system 1160 operates on or responds to data collected or generated by other systems of adaptive intelligence system 614, such as data processing systems, etc. It may consist of data processing, artificial intelligence, and computational systems 634 that may operate cooperatively to provide intelligence. In embodiments, providing coordinated intelligence may involve processing any of the demand management application output and supply chain application output as described in this application and the documents incorporated herein by reference to provide coordinated intelligence. It may include operating part of a set of artificial intelligence systems 1160 that utilize neural networks.

In embodiments, providing coordinated intelligence for a set of demand management applications 824 includes demand planning applications, demand forecasting applications, sales applications, future demand aggregation applications, marketing applications, advertising applications, e-commerce applications, and marketing analytics applications. , including customer relationship management applications, search engine optimization applications, sales management applications, advertising network applications, behavior tracking applications, marketing analytics applications, location-based product or service targeting applications, collaborative filtering applications, recommendation engines for products or services, etc. This may include configuring at least one of the adaptive intelligence systems 614 (e.g., via the user interface 3020, etc.) for at least one demand management application selected from a list of demand management applications.

Similarly, providing coordinated intelligence for a set of supply chain applications 812 may include product timing management applications, product quantity management applications, logistics management applications, shipping applications, delivery applications, order to product management applications, and component management applications. It may include configuring at least one of the adaptive intelligence system 614 for at least one supply chain application selected from a list of supply chain applications including orders for, etc.

In an embodiment, management platform 102 provides coordinated intelligence for a set of demand management applications 824 and supply chain applications 812 through the application of artificial intelligence, for example, through user interface 3020. May facilitate access to a set of adaptive intelligence systems 614. In this embodiment, the user may attempt to align supply and demand while ensuring the profitability of the value chain for the category 3010 of the product. By providing access to artificial intelligence capabilities (1160), the management platform allows users to benefit from technologies such as expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, etc., while enabling the management of supply and demand. Allows you to focus on your application.

In embodiments, management platform 102 may, through user interface 3020 or the like, perform demand management and supply chain management based, for example, on inputs used by applications, results generated by applications, and value chain results. An adaptive intelligence system that provides coordinated artificial intelligence 1160 for a set of demand management applications 824 and supply chain applications 812 for categories of goods 3020 by (automatically) determining relationships between the applications. A set of (614) can be provided. Artificial intelligence 1160 may be coordinated by a set of data processing, artificial intelligence, and computational systems 634 available through, for example, adaptive intelligence system 614.

In an embodiment, the management platform 102 includes a set of adaptive intelligence systems 614 that provide coordinated intelligence for a set of demand management applications 824 and supply chain applications 812 for categories of products 3010. It may consist of a set of artificial intelligence systems 1160 as part of . The set of artificial intelligence systems 1160 may cause at least one supply chain application in the set of supply chain applications 812 to select a product within a category of goods as determined by at least one demand management application in the set of demand management applications 824. Provide coordinated intelligence to produce results that address at least one aspect of provision for at least one of the In an example, a behavior tracking demand management application may generate results about the usage behavior of products within a category of products 3010. Artificial intelligence system 1160 may process the behavioral data and conclude that there is a perceived need for greater consumer accessibility to a second product within category of goods 3010. This coordinated intelligence can be selectively and automatically applied to a set of supply chain applications 812, for example, such that production resources or other resources in the value chain for a category of goods are allocated to a second product. In an example, a distributor handling retailer clearance receives a new inventory plan that allocates more retailer shelf space to a second product, e.g., by taking space away from a lower margin product, etc. can do.

In embodiments, a set of artificial intelligence systems 1160, etc. may affect control of a supply chain application such that, for example, selective temporal demand for at least one of the products within a category of products 3010 can be satisfied. It can provide coordinated intelligence for a set of supply chain and demand management applications by determining optional temporal prioritization of demand management application output. Seasonal adjustments in prioritization of demand application outcomes are one example of temporal variation. Adjustments to prioritization can also be localized, for example, when a large college football team is playing in its home stadium, tailgating (or tailgating) even though demand management application results suggest that small propane stoves are not currently in demand in the larger area. Local supply of tailgating supplies may be temporarily adjusted.

The set of adaptive intelligence systems 614 that provide coordinated intelligence, for example, by providing artificial intelligence capabilities 1160, may also provide automation of supply chain applications, automation of demand management applications, machine learning, artificial intelligence, intelligent transactions, etc. , may facilitate the development and deployment of intelligence for at least one function selected from a list of functions consisting of intelligent operation, remote control, analysis, monitoring, reporting, status management, event management, and process management. The set of adaptive intelligence systems 614 may be organized as a layer in the platform, with artificial intelligence systems therein being connected to other systems in the adaptive intelligence system layer (e.g., data processing systems, expert systems, machine learning systems, etc.). may act on or respond to data collected and/or generated by.

In addition to providing coordinated intelligence configured for a specific category of goods, coordinated intelligence may be provided for specific value chain entities 652, such as supply chain operators, businesses, enterprises, etc., participating in the supply chain for a category of goods. You can.

Providing coordinated intelligence may include using neural networks to process at least one of the inputs and outputs of a set of demand management and supply chain applications. Neural networks can be used in demand planning applications, demand forecasting applications, sales applications, future demand aggregation applications, marketing applications, advertising applications, e-commerce applications, marketing analytics applications, customer relationship management applications, search engine optimization applications, sales management applications, advertising network applications, It can be used with demand applications such as behavior tracking applications, marketing analytics applications, location-based product or service targeting applications, collaborative filtering applications, recommendation engines for products or services, etc. Neural networks can also be used with supply chain applications, such as product timing management applications, product quantity management applications, logistics management applications, shipping applications, delivery applications, order to product management applications, order to component management applications, etc. Neural networks can be used to manage processes, bill of materials, weather, traffic, design specifications, customer complaint logs, customer reviews, Enterprise Resource Planning (ERP) systems, Customer Relationship Management (CRM) systems, Customer Experience Management (CEM) systems, and Service Lifecycle Management (SLM). Coordinated intelligence can be provided by processing data available from any of a plurality of value chain data sources for categories of goods, including without limitation ) systems, Product Lifecycle Management (PLM) systems, etc. Neural networks configured to provide coordinated intelligence may share adaptive capabilities with other adaptive intelligence systems 614, such as when these systems are configured in a topology that facilitates such shared adaptation. In embodiments, a neural network may facilitate provisioning available value chain/supply chain network resources for both a set of demand management applications and a set of supply chain applications. In embodiments, a neural network may provide coordinated intelligence to improve at least one of a list of outputs consisting of process outputs, application outputs, process results, application results, etc.

15, a management platform of an information technology system, such as a management platform for a value chain of goods and/or services, is shown as a block diagram of functional elements and representative interconnections. The management platform includes, inter alia, a user interface 3020 that provides a hybrid set of adaptive intelligence systems 614. A hybrid set of adaptive intelligence systems 614 together with a set of supply chain applications 812 for categories of goods 3010 that can be produced and sold through the value chain and a set of demand management applications 824 For use, for example, through the application of hybrid artificial intelligence system 3060, through the application of artificial intelligence, and optionally through one or more expert systems, machine learning systems, etc., to provide coordinated intelligence. Hybrid adaptive intelligence system 614 may deliver two types of artificial intelligence systems, Type A 3052 and Type B 3054, through a set of data processing, artificial intelligence, and computational systems 634. In an embodiment, hybrid adaptive intelligence system 614 may operate on a set of supply chain applications (e.g., demand management application 824 and supply chain application 812). Alternatively, they may be selectable and/or configurable through the user interface 3020 to operate in cooperation with them. Hybrid adaptive intelligence system 614 may include any of a variety of expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, and other systems described throughout this disclosure and in documents incorporated by reference. It may include a hybrid artificial intelligence system 3060 that may include at least two types of artificial intelligence capabilities including those of. The hybrid adaptive intelligence system 614 applies a first type of artificial intelligence system 1160 to a set of demand management applications 824 and a second type of artificial intelligence system 1160 to a set of supply chain applications 812. Can facilitate application to, wherein each of the first and second types of artificial intelligence systems 1160 provides coordinated intelligence for the operation of a value chain that produces at least one of the products within the category of products 3010. It can operate independently, cooperatively, and selectively in coordinated motion to provide.

In an embodiment, user interface 3020 takes input from selected data sources in the value chain (such as data sources used by the set of demand management applications 824 and/or the set of supply chain applications 812) and For example, configuring hybrid artificial intelligence system 3060 to feed at least one of two types of artificial intelligence systems within hybrid artificial intelligence system 3060 - described throughout this disclosure and in documents incorporated by reference herein. The type may enhance, control, improve, optimize, configure, adapt, or otherwise affect the value chain for the category of product 3010, including an interface for doing so. In embodiments, selected data sources in the value chain may be applied as input for classification or prediction, or as results related to the value chain, categories of products 3010, etc.

In an embodiment, hybrid adaptive intelligence system 614 provides a plurality of distinct artificial intelligence systems 1160, hybrid artificial intelligence systems 3060, and combinations thereof. In embodiments, any of the plurality of distinct artificial intelligence systems 1160 and hybrid artificial intelligence systems 3060 may be configured as multiple neural network-based systems, such as classification adaptive neural networks, prediction adaptive neural networks, etc. As an example of a hybrid adaptive intelligence system 614, a machine learning-based artificial intelligence system may be provided for a set of demand management applications 824, and a neural network-based artificial intelligence system may be provided for a set of supply chain applications 812. can be provided. As an example of a hybrid artificial intelligence system 3060, hybrid adaptive intelligence system 614 may include a first type of artificial intelligence applied to a demand management application 824 and a second type applied to a supply chain application 812. A hybrid artificial intelligence system 3060 that is distinct from two types of artificial intelligence can be provided. Hybrid artificial intelligence system 3060 may include a plurality of first types of artificial intelligence (e.g., neural networks) and at least one second type of artificial intelligence (e.g., expert systems), etc. It may include any combination of systems. In an embodiment, a hybrid artificial intelligence system may include a hybrid neural network that applies a first type of neural network for a demand management application (824) and a second type of neural network for a supply chain application (812). Additionally, the hybrid artificial intelligence system 3060 may be used to support different demand management applications 824 (e.g., sales management applications and demand forecasting applications) or different supply chain applications 812 (e.g., logistics control applications and production quality control applications). Two types of artificial intelligence can be provided for different applications, such as:

In embodiments, hybrid adaptive intelligence system 614 may be applied as a separate artificial intelligence capability to separate demand management applications 824. For example, coordinated intelligence through hybrid artificial intelligence capabilities can be applied to demand planning applications by feed-forward neural networks, to demand forecasting applications by machine learning systems, to sales applications by self-organizing neural networks, and to sales applications by radial basis function neural networks. basis function neural network to future demand aggregation applications, convolutional neural networks to marketing applications, recurrent neural networks to advertising applications, hierarchical neural networks to e-commerce applications, stochastic neural networks to marketing analysis applications, provided to a customer relationship management application by an associative neural network, etc.

16, a management platform of an information technology system, such as a management platform for a value chain of goods and/or services, is shown as a block diagram of functional elements and representative interconnections for providing a set of predictions 3070. The management platform includes, inter alia, a user interface 3020 that provides a set of adaptive intelligence systems 614. Adaptive intelligence system 614 is for use with a coordinated set of demand management applications 824 and supply chain applications 812 for categories of goods 3010 that can be produced and sold through the value chain. Provides a set of predictions 3070 through the application of intelligence, such as through the application of an artificial intelligence system 1160, and optionally through one or more expert systems, machine learning systems, etc. Adaptive intelligence system 614 may communicate the set of predictions 3070 through a set of data processing, artificial intelligence, and computational systems 634. In embodiments, adaptive intelligence systems 614 are selectable through user interface 3020 such that one or more of adaptive intelligence systems 614 can operate on or in collaboration with a coordinated set of value chain applications and /or configurable. Adaptive intelligence system 614 may include any of a variety of expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, and other systems described throughout this disclosure and in documents incorporated by reference. It may include an artificial intelligence system that provides artificial intelligence capabilities known to be associated with artificial intelligence, including. Adaptive intelligence system 614 may facilitate, for example, coordinating two sets of value chain applications, or at least coordinating at least one demand management application and at least one supply chain application from each set. By creating a set of predictions 3070, it can facilitate applying adapted intelligence capabilities to a coordinated set of demand management applications 824 and supply chain applications 812.

In an embodiment, the set of forecasts 3070 may be at least one of the impact on the supply chain application based on the current state of the coordinated demand management application, such as a forecast that demand for a product will decline earlier than previously expected. Contains one prediction. The set of predictions 3070 is at least one of the impacts on the demand management application based on the current state of the coordinated supply chain application, such as a prediction that a shortage of supply of a product will likely affect the measure of demand for the related product. The opposite may also be true, in that it includes predictions of . In an embodiment, set of forecasts 3070 is a set of forecasts for adjustments in supply required to satisfy demand. Other forecasts include at least one forecast of changes in demand affecting supply. Another prediction within the set of predictions predicts a change in supply affecting at least one of the set of demand management applications, such as a promotional application for at least one product within a category of products. A forecast of a set of forecasts may be as simple as establishing the probability that the supply of a product in a category of product will not satisfy the demand established by the demand setting application.

In embodiments, adaptive intelligence system 614 may provide a set of artificial intelligence capabilities to facilitate providing a set of predictions for a coordinated set of demand management applications and supply chain applications. In one non-limiting example, the set of artificial intelligence capabilities may include probabilistic neural networks that can be used to predict fault conditions or problem states in a demand management application, such as lack of sufficient verified feedback. Probabilistic neural networks are used to perform value chain operations on machines (e.g. production machines, automated handling machines, packaging machines, shipping machines, etc.) based on a collection of machine operation information and preventive maintenance information about the machines. ) can be used to predict the problem state.

In embodiments, the set of predictions 3070 may be provided directly by the management platform 102 through a set of adaptive artificial intelligence systems.

In embodiments, the forecast set 3070 may be provided for a coordinated set of demand management applications and supply chain applications for a category of goods by applying artificial intelligence capabilities to coordinate the set of demand management applications and supply chain applications. there is.

In an embodiment, the set of predictions 3070 may be predictions of outcomes for operating the value chain with a coordinated set of demand management applications and supply chain applications for a category of goods, such that the user can determine which set of outcomes is desirable. perform test cases on a coordinated set of demand management applications and supply chain applications to determine which sets may produce undesirable results (viable candidates for the coordinated set of applications) and which sets may produce undesirable results. You can.

17, a management platform of an information technology system, such as a management platform for a value chain of goods and/or services, is depicted as a block diagram of functional elements and representative interconnections to provide a set of classifications 3080. . The management platform includes, inter alia, a user interface 3020 that provides a set of adaptive intelligence systems 614. The adaptive intelligence system 614 provides a demand management application ( 824) and supply chain applications 812, for example, through the application of artificial intelligence, for example, through the application of artificial intelligence system 1160, and optionally one or more expert systems. , provides a set of classifications (3080) through a machine learning system, etc. Adaptive intelligence system 614 may convey the set of classifications 3080 through a set of data processing, artificial intelligence, and computational systems 634. In embodiments, adaptive intelligence systems 614 are selectable through user interface 3020 such that one or more of adaptive intelligence systems 614 can operate on or in concert with a coordinated set of value chain applications and /or configurable. Adaptive intelligence system 614 may, among other things, include various expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, and other systems described throughout this disclosure and in documents incorporated by reference. Any of these may include an artificial intelligence system that provides classification capabilities. Adaptive intelligence system 614 may facilitate, for example, coordinating two sets of value chain applications, or at least coordinating at least one demand management application and at least one supply chain application from each set. By creating a set of classifications 3080 that can facilitate applying adapted intelligence capabilities to a coordinated set of demand management applications 824 and supply chain applications 812.

In an embodiment, the classification set 3080 may represent at least one current state of the supply chain application for use by a coordinated demand management application, such as a classification of problem states that may impact the operation of a demand management application, such as a marketing application. Contains one category. This classification can be useful in determining how to adjust market expectations for products that will have lower yields than previously expected. The opposite may be true in that the set of classifications 3080 includes at least one classification of the current state of the demand management application and its relationship to the coordinated supply chain application. In an embodiment, set of categories 3080 is a set of categories of adjustments to supply required to meet demand, for example, adjustments to production worker demands would be classified differently than adjustments to third-party logistics providers. Other classifications may include at least one classification of perceived demand changes and the resulting potential impact on supply management. Another classification within the set of classifications may include supply chain application impact on at least one of the set of demand management applications, such as a promotional application for at least one product within a category of products. A classification within a set of classifications may be as simple as classifying the likelihood that the supply of goods within a category of goods will not satisfy the demand established by the demand setting application.

In embodiments, adaptive intelligence system 614 may provide a set of artificial intelligence capabilities to facilitate providing a set of classifications 3080 for a coordinated set of demand management applications and supply chain applications. In one non-limiting example, the set of artificial intelligence capabilities may include probabilistic neural networks that can be used to classify fault conditions or problem states in demand management applications, such as those lacking sufficient validated feedback. there is. A probabilistic neural network links the problem state of machines performing value chain operations (e.g., production machines, automated handling machines, packaging machines, delivery machines, etc.) to at least one of machine operation information and preventive maintenance information for the machines. Can be used to classify things as related.

In embodiments, the set of classifications 3080 may be provided directly by the management platform 102 through a set of adaptive artificial intelligence systems. Additionally, a set of classifications 3080 may be provided for a coordinated set of demand management applications and supply chain applications for a category of goods by applying artificial intelligence capabilities to coordinate the set of demand management applications and supply chain applications. .

In an embodiment, the set of classifications 3080 may be the resulting classifications for operating a value chain with a coordinated set of demand management applications and supply chain applications for a category of goods, such that the user may determine which set is desirable. of demand management applications and supply chain applications to generate results that are classified (e.g., viable candidates for a coordinated set of applications) and determine which sets may produce results that are classified as undesirable. A coordinated set of test cases can be performed.

In embodiments, the set of classifications may include a set of adaptive intelligence functions, such as a neural network, that can be adapted to classify information associated with a category of products. In an example, the neural network may be a multilayer feedforward neural network.

In embodiments, performing classification may include classifying the discovered value chain entities as either demand-driven or supply-driven.

In an embodiment, the set of classifications 3080 may be achieved through the use of an artificial intelligence system 1160 to coordinate a coordinated set of demand management and supply chain applications. The artificial intelligence system may configure and generate a set of classifications 3080 as a means by which demand management applications and supply chain applications may be coordinated. In one example, the classification of information flows across the value chain can be classified as relevant to both demand management applications and supply chain applications; These common associations can be points of coordination between applications. In an embodiment, the set of classifications may be artificial intelligence generated classifications resulting from operating a supply chain that relies on the coordinated demand management application 824 and supply chain application 812.

18, a management platform of an information technology system, such as a management platform for a value chain of goods and/or services, is shown as a block diagram of functional elements and representative interconnections to achieve automated control intelligence. The management platform includes, inter alia, a user interface 3020 that provides a set of adaptive intelligence systems 614. Adaptive intelligence system 614 provides automated control signaling ( 3092) is provided. Adaptive intelligence system 614 may convey automated control signals 3092 through a set of data processing, artificial intelligence, and computational systems 634. In an embodiment, adaptive intelligence system 614 may automatically configure a set of supply chain applications (e.g., demand management application 824 and supply chain application 812). It is selectable and/or configurable through user interface 3020 for control. Adaptive intelligence system 614 may include any of a variety of expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, and other systems described throughout this disclosure and in documents incorporated by reference. It may include artificial intelligence, including:

In embodiments, user interface 3020 may be used to configure selected data sources of value chain 3094 (e.g., data sources used by a coordinated set of demand management applications 824 and/or a set of supply chain applications 812). Automated control signals to take inputs from and effect them, e.g., to enhance, control, improve, optimize, configure, adapt, or otherwise affect the value chain for a category of goods 3010. Adaptable to feed any of the neural networks, artificial intelligence system 1160, or other adaptive intelligence systems 614, such as those described throughout this disclosure and in documents incorporated by reference herein, to generate 3092. It may include an interface for configuring the intelligent system 614. In embodiments, selected data sources in the value chain may be used to determine aspects of automated control signals, such as temporal adjustments to control outcomes associated with the value chain, at least for categories of goods 3010, etc.

In one example, the set of automated control signals is at least one for automating the execution of supply chain applications, such as production start, automated material ordering, inventory verification, billing applications, etc., in a coordinated set of demand management applications and supply chain applications. May include control signals. In another example of automated control signal generation, the set of automated control signals is at least one for automating the execution of a demand management application, such as a product recall application, an email distribution application, etc., in a coordinated set of demand management applications and supply chain applications. It may include control signals. In another example, automated control signals may control the timing of a demand management application based on product supply status.

In embodiments, adaptive intelligence system 614 may apply machine learning to the results of supply to automatically adapt a set of demand management application control signals. Similarly, adaptive intelligence system 614 may apply machine learning to the results of demand management to automatically adapt a set of supply chain application control signals. Adaptive intelligence system 614 can automate, for example, by applying artificial intelligence to determine aspects of the value chain that affect the automated control of a coordinated set of demand management applications and supply chain applications for a category of goods. Additional processing can be provided to generate the control signals. The determined aspects can be used in the generation and operation of automated control intelligence/signals, such as by filtering value chain information for aspects that do not impact target demand management and supply chain applications.

For example, automated control of a supply chain application may be limited by, for example, policies, operational restrictions, safety constraints, etc. A set of adaptive intelligence systems can determine the range of supply chain application control values over which controls can be automated. In embodiments, the range may be associated with supply rate, supply timing rate, mix of products within a category of products, etc.

Embodiments for using artificial intelligence systems or capabilities to identify, configure, and condition automated control signals are described herein. Such embodiments include demand management and supply chain applications that are selectively processed with machine learning and used to adapt automated control signals for at least one of the products within a category of products (e.g., status information, output information, results, etc. ) may further include a closed loop of feedback from the adjusted set. Automated control signals may be adapted, for example, based on indications of feedback from a supply chain application that the yield of goods suggests a production problem. In this example, an automated control signal can affect production speed, and feedback can cause the signal to automatically self-adjust to a slower production speed until the production problem is resolved.

19, a management platform of an information technology system, such as a management platform for a value chain of goods and/or services, is shown as a block diagram of functional elements and representative interconnections for providing information routing recommendations. The management platform includes a set of value chain networks 3102 where network data 3110 is collected from a set of information routing activities, and the information includes results, parameters, routing activity information, etc. Within the set of value chain networks 3102, a selection value chain network 3104 is selected for which at least one information routing recommendation 3130 is provided. Artificial intelligence system 1160 may include a machine learning system and may be trained using a training set derived from network data 3110 results, parameters, and routing activity information for a set of value chain networks 3102. . The artificial intelligence system 1160 may further provide information routing recommendations 3130 based on the current state 3120 of the selected value chain network 3104. The artificial intelligence system may use machine learning to train on information transaction types within the set of value chain networks 3102, thereby enabling different transaction types (e.g., real-time inventory updates, buyer credit checks, engineering sign-offs) signoff, etc.) and contribute to information routing recommendations accordingly. The artificial intelligence system may also use machine learning to train about the information value of different types and/or classes of information routed within and across the set of value chain networks 3102. Information is important not only for the timing of information availability and the timing of information consumption, but also for value chain network elements (e.g. production providers) to take desired actions (e.g. starting mass production without work orders) without that information. Values can be placed on a wide range of factors, including values based on information content, such as information that cannot be valued. Accordingly, information routing recommendations may be based on training on transaction type, information value, and combinations thereof. These are merely exemplary information routing recommendation training and recommendation basis factors and are presented herein without limitation to other elements for the training and recommendation basis.

In embodiments, artificial intelligence system 1160 may provide information routing recommendations 3130 based on transaction type, transaction type and information type, network type, etc. Information routing recommendations may be based on a combination of factors such as information type and network type, such as when the information type (streaming) is incompatible with the network type (small transaction).

In embodiments, artificial intelligence system 1160 may use machine learning to develop an understanding of networks within selected value chain network 3104, such as network topology, network loading, network reliability, network latency, etc. This understanding, for example, can be combined with detected or expected network conditions to form informative routing recommendations. Aspects such as the presence of edge intelligence in the value chain network 3104 may influence one or more information routing recommendations. In one example, the type of information may be incompatible with the network type; However, the network may be comprised of edge intelligence that can be utilized by the artificial intelligence system 1160 to adapt the type of information being routed, so that it is compatible with the targeted network type. This is also an example of information routing recommendations such as edge computing, server access, network-based storage resources, etc. - more general considerations for network resources (e.g., presence, availability, and capability). Likewise, value chain network entities can influence information routing recommendations. In embodiments, information routing recommendations may avoid routing information that is confidential to a first supplier in the value chain through a network node controlled by a competitor of the supplier. In embodiments, information routing recommendations route information to a first node where the information is partially consumed and partially processed for further routing, such as splitting the partially processed portion into destination-specific information sets. It may include:

In embodiments, artificial intelligence system 1160 may provide information routing recommendations based on goals, such as goals of the value chain network, goals of information routing, etc. Goal-based information routing recommendations may include routing objectives such as quality of service routing objectives, routing reliability objectives (which may be measured based on transmission failure rate, etc.). Other goals may include a measure of latency associated with one or more candidate paths. Information routing recommendations may be based on the availability of information in the selected value chain network, such as when the information is available and when it needs to be delivered. For information that is available well in advance of the time it is needed (for example, an overnight production report that becomes available for routing at 2 a.m. will be needed first at 7 a.m.), routing recommendations may involve short delays in routing, etc. This may include using lower cost resources available. For information that is obtained shortly before the point of need (for example, the results of a product test are needed within a few hundred milliseconds of the test being completed to maintain production run speed, etc.).

Information routing recommendations may be formed by the artificial intelligence system 1160 based on information persistence factors, such as how long the information is available for immediate routing within the value chain network. Information routing recommendations that consider information persistence can select network resources based on availability, cost, etc. during the time of information persistence.

Information value and its impact on information value can be considered in information routing recommendations. By way of example, information valid for a single shipment (e.g., one production run of goods) may lose substantial value once the shipment is satisfactorily received. In this example, an information routing recommendation may refer to routing relevant information to all of the highest priority consumers of that information while the relevant information is still valid. Likewise, the routing of information consumed by more than one value chain entity can be done at the desired time/moment of each value chain entity, e.g. during the same production shift, at the beginning of their day - if the entities are in different time zones. The back may be different - the back may need to be adjusted to receive information.

In embodiments, information routing recommendations may be based on the topology of the value chain, location and availability of network storage resources, etc.

In embodiments, while information is being routed, for example, based on changes in network resource availability, network resource discovery, network dynamic loading, priority of recommendations generated after the information for the first recommendation is within the path, etc. , one or more information routing recommendations may be adapted.

Referring to FIG. 20, a management platform of an information technology system, such as a management platform for the value chain of goods and/or services, includes functional elements for semi-perceptual problem recognition of pain points in the value chain network, and It is shown as a block diagram of representative interconnections. The management platform includes a set of value chain network entities 3152 from which entity-related data 3160 is collected, including results, parameters, activity information, etc. associated with the entities. Within the set of value chain network entities 3152, a set of select value chain network entities 3154 for which at least one pain point condition 3172 is detected is selected. The artificial intelligence system 1160 is trained on a training set derived from entity-related data 3160 that includes training on results associated with value chain entities, e.g., parameters associated with the operation of the value chain, value chain activity information, etc. It can be. The artificial intelligence system may further utilize machine learning to facilitate learning problem state factors 3180 that can characterize the problem state input as training data. These factors 3180 may be further utilized by an instance of artificial intelligence 1160' operating on computing resources 3170 local to the value chain network entity experiencing the pain points/consequences. The goal of this configuration of artificial intelligence systems, data sets, and value chain networks is to recognize problem states in selected parts of the value chain.

In embodiments, recognizing problem states may be based on analysis of fluctuations, such as fluctuations occurring in value chain metrics (e.g., loading, latency, delivery time, cost, etc.), particularly at certain scales over time. there is. Variations that exceed variation thresholds (e.g., selective dynamic range of outcomes of value chain operations such as production, delivery, customs clearance, etc.) may indicate pain points.

In addition to detecting problem states, platform 102 predicts pain points based at least in part on correlations with detected problem states, such as through methods of semi-perceptual problem recognition. These correlations can be derived from the value chain, such as a shipper not being able to deliver international goods until the goods are cleared through customs, or sales forecasts not being provided with high confidence without high-quality field data. In embodiments, predicted pain points may be points of additional value chain activities along the supply chain, activities arising from related activities (e.g., tax planning related to tax law), etc. Predicted pain points may be assigned a risk value based on the aspect of the problem state detected and the correlation between the predicted pain point activity and the problem state activity. If a production operation may receive material from two suppliers, a problem condition from one of the suppliers may indicate a pain point in the use of low-risk material. Likewise, if a demand management application indicates high demand for a product and detects a problem with the information on which that demand is based, for example, overstock (a pain point) depending on how far the product has progressed in the value chain. The risk may be high.

In embodiments, semi-perceptual problem recognition may involve more than simple links of data and operational states of entities involved in the value chain. Problem recognition may also be based on human factors, such as the perceived stress of production supervisors, deliverers, etc. Human factors for use in semi-perceptual problem recognition may be collected from sensors (e.g., wearable physiological sensors, etc.) that facilitate detection of human stress levels, etc.

In embodiments, semi-perceptual problem recognition may also be based on unstructured information, such as digital communications, voice messaging, etc., that may be shared between, sent with, or received by humans involved in value chain operations. can do. As an example, natural language processing of email communications between workers within an enterprise may indicate, for example, a level of inconvenience to suppliers up the value chain. Although data associated with a supplier (e.g., on-time production, quality, etc.) may be within the range of variation considered acceptable, information within this unstructured content may be associated with potential pain points, such as personal issues with the supplier's key participants. can indicate. By using natural language processing, artificial intelligence, and optionally machine learning, problem state recognition can be improved.

In embodiments, semi-perceptual problem recognition may be based on analysis of changes in the scale of a value chain operation/entity/application, including changes in a given measure over time, changes in two related measures, etc. In embodiments, variation in results over time may indicate problem status and/or suggest pain points. In embodiments, an artificial intelligence-based system determines an acceptable range of outcome variation and applies that range to measurements of a selected set of value chain network entities, such as entities that share one or more similarities, to facilitate detection of problematic conditions. can do. In an embodiment, the acceptable range of outcome variation may indicate a problem state trigger threshold that can be used by a local instance of artificial intelligence for signaling the problem state. In this scenario, a problem state may be detected when at least one metric of a value chain activity/entity, etc. is greater than an artificial intelligence determined problem state threshold. Variation analysis for problem state detection may include variations in at least one of the following: variation in start/end times of scheduled value chain network entity activities, production time, production quality, production rate, production start time, production resource availability, or trends thereof, delivery, Variations in the measurement of supply chain entities, variations in the duration of time for transfer from one mode of transport to another (e.g. when the variation is greater than the transport mode problem status threshold), variations in quality testing, etc. It may include detecting.

In embodiments, the semi-sentient problem recognition system may further respond along the supply chain due to detected pain points, such as risk and/or need for overtime, expedited delivery, product price discounts, etc. May include machine learning/artificial intelligence prediction of identified pain points.

In embodiments, the machine learning/artificial intelligence system may be used to: late delivery, damaged container, damaged goods, incorrect goods, customs delays, unpaid duties, weather events, damaged infrastructure, blocked waterways, incompatible infrastructure, congested ports, Congested handling infrastructure, congested roads, congested distribution centers, rejected goods, returned goods, waste, wasted energy, wasted labor, untrained labor, poor customer service, empty transport vehicles on return routes; Process results, parameters, and data collected from a set of data sources associated with a set of value chain entities and activities to detect at least one pain point selected from a list of pain points consisting of excessive fuel prices, excessive tariffs, etc. there is.

21, a management platform of an information technology system, such as a management platform for a value chain of goods and/or services, is an automated coordination of functional elements and representative interconnections of a set of value chain network activities for a set of products of an enterprise. It is shown as a block diagram of . The management platform provides activity information 3208 used by the artificial intelligence system 1160 to provide automated coordination 3220 of value chain network activities 3212 for a set of products 3210 for an enterprise 3204. It includes a set of network-connected value chain network entities 3202 that create a network. In embodiments, value chain monitoring system 614 monitors the activity of a set of networked value chain entities 3202 and works cooperatively with data collection and management system 640 to collect value chain information, such as activity information, configuration information, etc. Entity monitoring information can be collected and stored. This collected information may be organized as activity information 3208 for a set of activities associated with a set of products 3210 of an enterprise 3204. In embodiments, artificial intelligence system 1160 may use application programming connectivity facility 642 to automate access to monitored activity information 3208.

A value chain can include multiple interconnected entities, each performing several activities to complete the value chain. Although humans play an important role in some activities within the value chain network, artificial intelligence-type systems (including, for example, such systems described in this application and the documents incorporated herein by reference) are used to coordinate supply chain activities. Greater automated coordination and integrated orchestration of supply and demand can be achieved using machine learning, expert systems, self-organizing systems, etc. The use of artificial intelligence not only provides greater capabilities to end users, but also the emerging nature of self-adaptive systems, including Internet of Things (IoT) devices and intelligent products, which can play a key role in the automated coordination of supply chain activities. can be made more abundant.

For example, an IoT system deployed in a fulfillment center 628 may be coordinated with an intelligent product 1510 that takes customer feedback regarding a product 1510 and may provide applications for the fulfillment center 628 ( Upon receiving customer feedback regarding a problem with product 1510 through a connection path by intelligent product 1510, 630 may request a similar product 650 before product 650 is dispatched from fulfillment center 628. ) can initiate a workflow to perform corrective action. Workflow analyzes issues with product 1510, develops an understanding of the value chain network activities that produce the product, determines the resources required for the workflow, and conducts inventory and production operations to adapt any existing workflow. It may be configured by an artificial intelligence system 1160 that coordinates with the system, etc. Artificial intelligence system 1160 may further coordinate with demand management applications to address any temporary effects on product availability, etc.

In embodiments, automated coordination of a set of value chain network activities for a set of products for an enterprise, e.g., to provide selectively coordinated intelligence and coordinate activities, such as demand management activities, supply chain activities, etc. may rely on the methods and systems of coordinated intelligence described herein to facilitate coordinating . As an example, artificial intelligence can facilitate determining relationships between value-changing network activities based on the inputs used by the activities and the outcomes generated by the activities. Artificial intelligence can be used by value chain network entities to continuously monitor activities, identify temporal aspects that require adjustments (for example, when changes in supply have a temporal impact on demand activity), and automate such adjustments. Integrate with and/or work in collaboration with activities on platforms such as Activities. An advanced artificial intelligence system that can enable the use of different artificial intelligence capabilities for any given value chain set of entities, applications, or conditions. You can benefit from The use of hybrid artificial intelligence systems can provide advantages by applying more than one type of intelligence to a set of conditions to facilitate human and/or computer automated selection. Artificial intelligence further enhances the automated coordination of value chain network entity activities through intelligent actions such as generation of sets of predictions, sets of classifications, generation of automated control signals (which can be communicated across value chain network entities, etc.) You can do it. Other exemplary artificial intelligence-based impacts on automated coordination of value chain network entity activities include machine learning-based information routing and recommendations therefor, structured (e.g., production data) and unstructured (e.g., human emotions) s) includes semi-perceptual problem recognition based on both sources. The artificial intelligence system will facilitate the automated coordination of the activities of value chain network entities for a set of products or an enterprise based on the adaptive intelligence provided by the platform for the categories of goods into which the enterprise's set of products can be grouped. You can. In an example, adaptive intelligence may be provided by a platform for a drapery hanging category of merchandise, and a set of products for a business may include a line of adaptable drapery hangers. With the understanding developed for the entire drapery hanging category, artificial intelligence capabilities can be applied to a company's value chain network activities to automate aspects of the value chain, such as the exchange of information between activities.

DIGITAL TWIN SYSTEM IN VALUE CHAIN ENTITY MANAGEMENT PLATFORM

22, the adaptive intelligence layer 614 may include a value chain network digital twin system 1700, for visualization of various value chain entities 652, environments, and applications 630. as well as digital twin capabilities for coordinated intelligence and other value-added services and capabilities (including artificial intelligence (1160), edge intelligence (1400), analytics and other capabilities) enabled or facilitated by the digital twin (1700). may include a set of components, processes, services, interfaces, and other elements for the development and deployment of Without limitation, digital twin 1700 may be used and/or applied to each of the processes managed, controlled, or mediated by each of the set of applications 614 in the platform application layer.

In embodiments, digital twin 1700 may take advantage of the presence of multiple applications 630 within value chain management platform 604, such that a pair of applications can be used as data sources ( e.g., from the data storage layer 624) and other inputs (e.g., from the monitoring layer 614), as well as artificial intelligence 1160 (a variety of expert systems, artificial intelligence systems, neural networks, supervised learning, through the use of systems, machine learning systems, deep learning systems, and other systems described throughout this disclosure and in documents incorporated by reference) and a monitoring layer 614 and data collection system ( Outputs, events, state information, and outputs can be shared, including through the use of content collected by 640), which collectively provide a much richer environment for enriching content in the digital twin 1700. You can.

In embodiments, digital twin 1700 involves shared or converged processes between various pairs of applications 630 of applications 604, such as, but not limited to, security applications 834 and inventory management applications 820. It can be used in connection with converged processes, integrated automation of facility management applications 850 and blockchain-based applications 844, etc. In an embodiment, the converged process may include multiple applications 630 that can be coupled with the digital twin 1700 such that the digital twin 1700 is updated accordingly (tracking the same transaction on the blockchain, but of data objects maintained on the blockchain). may include shared data structures (including those that can consume different subsets of available attributes or use a set of nodes and links in a common knowledge graph). For example, a transaction representing a change in ownership of entity 652 may be stored in a blockchain and, for example, linked to and shared with digital twin 1700 so that digital twin 1700 can be updated accordingly. It can be used by multiple applications 630 to enable role-based access control, role-based permissions for remote control, identity-based event reporting, etc. In an embodiment, the converged process is a larger flow that involves one or more of the set of applications 614 that can be connected to and shared with digital twin 1700 so that digital twin 1700 can be updated accordingly. May include shared process flows across applications 630, including subsets. For example, an inspection flow for a value chain network entity 652 may serve an analytics solution 838, an asset management solution 814, etc.

In embodiments, digital twin 1700 may be provided for a wide range of value chain network applications 630 mentioned throughout this disclosure and documents incorporated by reference herein. The environment for development of the digital twin 1700 allows developers to take input from selected data sources in the data storage layer 624 and events or other data from the monitoring system layer 614 and include them in the digital twin 1700. It may include a set of interfaces for developers to configure the artificial intelligence system 1160 to provide them. The digital twin 1700 development environment can be configured to take output and results from various applications 630.

VALUE CHAIN NETWORK DIGITAL TWINS

23 , as described throughout this disclosure, a component or element of the value chain network entity 652, application 630, or platform 604, e.g., event data 1034, status, etc. By populating digital twin 1700 with value chain network data objects 1004, such as data 1140 or other data, any of the value chain network entities 652 are represented in a set of one or more digital twins 1700. It can be.

Accordingly, platform 604 may include, integrate, be integrated with, manage, control, coordinate, or otherwise handle any of the following, among others: distribution twin 1714 (e.g., A wide variety of digital twins (1700), such as those representing distribution facilities, assets, objects, workers, etc.; Warehousing Twin (1712) (e.g., representing warehouse facilities, assets, goods, workers, etc.); port infrastructure twin 1714 (e.g., representing a port, airport, or other facility, as well as assets, objects, workers, etc.); Shipping Facility Twin (1720); Operating Facility Twin (1722); customer twin 1730 (e.g., representing the physical, behavioral, demographic, psychological, financial, historical, affinity, interests, and other characteristics of a group of customers or an individual customer); Worker twin 1740 (e.g., physical attributes, physiological data, state data, psychological information, emotional state, state of fatigue/energy, state of attention, skills, training, competencies, roles, authority, responsibilities, work status, activities , and other properties of or accompanying the worker); Wearable/Portable Device Twin (1750); Process Twin (1760); machine twin 21010 (e.g., for the various machines used to support the value chain network 668); product twin (1780); Origin point twin (1560); provider twin (1630); supplyfactor twin(1650); Marine Facility Twin (1572); Floating Asset Twin (1570); Shipyard Twin (1620); Destination Twin (1562); Fulfillment Twin (1600); Delivery System Twin (1610); demand factor twin (1640); Retailer Twin (1790); E-commerce and online site and operator Twin (1800); Aqueduct Twin (1810); Road Twin (1820); Railway Twin (1830); aviation facility twins 1840 (e.g., twins of aircraft, runways, airports, hangars, warehouses, air travel corridors, refueling facilities, and other assets, objects, personnel, etc. used in connection with the air transportation of products 650); Autonomous Vehicle Twin (1850); Robotics Twin (1860); Drone Twin (1870); and Logistic Factor Twin (1880); etc. Each of these includes: mirroring or reflecting changes in the state of an associated physical object or other entity, providing the ability to model the behavior or interaction of an associated physical object or other entity, enabling simulation, state The digital twin may have the characteristics described throughout this disclosure and the documents incorporated by reference in this application, such as providing a representation of.

In an example embodiment, a digital twin system may be configured to create various enterprise digital twins 1700 associated with a value chain (e.g., specifically value chain network entity 652). For example, a company that produces goods internationally (or at multiple facilities) might have supplier twins representing the company's supply chain, factory twins at various production facilities, product twins representing the products made by the company, and distribution twins for the company. A set of digital twins 1700 may be constructed, such as a distribution twin representing the chain, and other suitable twins. In doing so, companies can define the structural elements of each individual digital twin, as well as arbitrary system data that corresponds to the structural elements of the digital twin. For example, in creating a production facility twin, a company can shape the layout and spatial definition of the facility and any processes performed in the facility. Enterprises can also define data sources that correspond to value chain network entities 652, such as sensor systems, smart manufacturing equipment, inventory systems, logistics systems, etc., that provide data related to facilities. Companies can associate data sources with elements of a production facility and/or processes that occur in the facility. Similarly, enterprises can define the structural, process, and layout definitions of their supply chains and distribution chains, and connect relevant data sources such as supplier databases and logistics platforms to create respective distribution chains and supply chain twins. . Companies can further relate these digital twins to have a view of the value chain. In embodiments, a digital twin system may perform a simulation of an enterprise's value chain incorporating real-time data obtained from the enterprise's various value chain network entities 652. In some of these embodiments, the digital twin system can determine when to order specific parts to manufacture a specific product, when to schedule maintenance on a machine, and/or when to install a machine, given the predicted demand for the manufactured product. When to ship the item (for example, when a digital simulation of the digital twin indicates there may be the least demand for a particular product or when it will have the least impact on the company's income statement); Decisions such as the like may be recommended to users interacting with the enterprise digital twin 1700. The foregoing example is a non-limiting example of how a digital twin can collect system data and perform simulations to add one or more objectives.

ENTITY DISCOVERY AND INTERACTION MANAGEMENT

24, a monitoring system layer (e.g., IoT data collection system, data collection system that searches social networks, websites, and other online resources, crowdsourcing system, etc.) including various data collection systems 640 ( 614), as well as for identifying a set of value chain network entities 652, identifying a type of value chain network entity 652, identifying a specific value chain network entity 652, etc. , also for resolving ambiguities (e.g., when a single entity is identified differently in different systems, when different entities are identified similarly, etc.), for entity identity deduplication, and for entity identity analysis. Entity discovery system 1900, such as for managing the identity of value chain network entities 652, including for identity enhancement (e.g., by enriching data objects with additional data collected about entities within the platform), etc. ) may include a set of Entity discovery 1900 also determines how entities are connected (e.g., by what network connections, data integration systems, and/or interfaces) and what data is exchanged between entities (what types of data objects are exchanged). , what common workflows involve the entities, what inputs and outputs are exchanged between the entities, etc.), what rules or policies govern the entities, etc. there is. Platform 604 discovers through entity discovery 1900, including learning based on a training data set to manage interactions between entities based on how the entities have been managed by human supervisors or other systems. Includes a set of entity interaction management systems 1902, which may include one or more artificial intelligence systems (including any of the types described throughout this disclosure) for managing a set of interactions between entities. can do.

As an illustrative example among many possibilities, the entity discovery system 1900 could discover network-connected cameras showing the shipping docks of a facility that produces products for an enterprise, as well as determine which interface to access the feed of video content from the cameras. Alternatively, it can be used to identify whether a protocol is needed. The entity interaction management system 1902 may then establish access to the feed and provide the feed to other systems for further processing, such as the artificial intelligence system 1160 processing the feed to discover content relevant to the enterprise's activities. It can be used to interact with an interface or protocol to do something. For example, artificial intelligence system 1160 may create markings (e.g., labels, SKUs, images, logos, etc.) that may indicate that a product has moved through a shipping dock, shapes (e.g., packages of a particular size or shape), image frames of the video feed can be processed to find activities (e.g., loading or unloading activities), etc. This information can be used to replace, augment, or verify other information, such as RFID tracking information. Similar discovery and interaction management activities may be performed on any of the types of value chain network entities 652 described throughout this disclosure.

ROBOTIC PROCESS AUTOMATION IN VALUE CHAIN NETWORK

25, the adaptive intelligence layer 614 is a layer of components, processes, services, interfaces and other elements for the development and deployment of automation capabilities for various value chain entities 652, environments, and applications 630. A robotic process automation (RPA) system 1442 may include a set. Without limitation, robotic process automation 1442 may include, but is not limited to, each of the processes managed, controlled, or mediated by each of the set of applications 614 of the platform application layer, functions, components, workflows, and processes of VCNP 604 itself; It can be applied to processes involving value chain network entities 652 and other processes.

In embodiments, robotic process automation 1442 may take advantage of the presence of multiple applications 630 within value chain management platform 604, such that a pair of applications may be collected in association with value chain entity 652. Not only can data sources (e.g., from the data storage layer 624) and other inputs (e.g., from the monitoring layer 614) be shared, but also (as described throughout this disclosure and in documents incorporated by reference) Outputs, events, states, including through the use of artificial intelligence 1160 (including any of a variety of expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, and other systems) Information and output can be shared, and they can collectively provide a much richer environment for process automation. For example, the asset management application 814 may be able to perform a human-based inspection (e.g., by automating processes involving visual inspection using video or still images, such as from a camera displaying an image of the entity 652). Robotic process automation 1442 may be used for the automation of asset inspection processes normally performed or supervised by, e.g., where the robotic process automation 1442 system identifies, diagnoses, and measures possible defects or favorable characteristics of a facility or other asset. , is trained to automate inspection by observing the interaction of a set of human inspectors or supervisors with the interfaces used to parameterize, or otherwise characterize. In embodiments, the interaction of a human inspector or supervisor may include a labeled data set where the label or tag indicates a type of defect, a favorable characteristic, or another characteristic, such that the machine learning system can identify the same characteristic. can be learned using training data sets, which in turn can be used to automate inspection processes such that defects or favorable characteristics are automatically classified and detected in sets of video or still images, which in turn can be used in value chain network asset management applications. 814) can be used to flag items as requiring further inspection, to be rejected, to be disclosed to prospective purchasers, to be corrected, etc. In embodiments, robotic process automation 1442 may involve multi-application or cross-application sharing of inputs, data structures, data sources, events, states, outputs, or results. For example, asset management application 814 may enrich robotic process automation 1442 of asset management application 814, such as with information regarding the current characteristics of an item from a specific vendor in the supply chain for an asset. may receive information from a marketplace application 854, which may assist in populating characteristics regarding the asset for the purpose of facilitating the inspection process, negotiation process, delivery process, etc. These and many other examples of multi-application or cross-application sharing for robotic process automation 1442 across applications 630 are included in this disclosure. Robotic process automation 1442 may be used with various functions of VCNP 604. For example, in some embodiments, robotic process automation 1442 may be described as training robots to operate and automate tasks that were, at least largely, handled by humans. One of these tasks can be used to train a robot that can then train other robots. Robotic process automation 1442 can be trained (e.g., through machine learning) to mimic interactions on a training set, and these trained robotic process automations 1442 (e.g., trained agents Alternatively, trained robotic process automation systems) can perform these tasks previously performed by humans. For example, robotic process automation 1442 may observe software interactions (e.g., mouse movements, Software is available that can provide mouse clicks, cursor movements, navigation actions, menu selections, keyboard typing, etc.). This may include monitoring the user's mouse clicks, mouse movements, and/or keyboard typing and learning to perform the same clicks and/or typing. In another example, robotic process automation 1442 may utilize software to learn robots and physical interactions with other systems to train the robotic system to sequence or perform the same physical interactions. For example, a robot could be trained to rebuild a set of bearings by having the robot watch a video of someone performing this task. This may include tracking physical interactions and tracking interactions at a software level. Robotic process automation 1442 may understand what underlying capabilities the VCNP 604 is positioned to pre-construct a combination of neural networks that can be used to replicate the performance of human capabilities.

In embodiments, robotic process automation may include, but is not limited to, converged processes involving security applications 834 and inventory applications 820, integrated automation of blockchain-based applications 844 with vendor management applications 832, etc. The same may apply to shared or converged processes between various pairs of applications 630 of application 604 . In embodiments, the converged process may be a shared data structure for multiple applications 630 (one that tracks the same transaction on the blockchain but may consume different subsets of the available properties of the data objects maintained on the blockchain). including using a set of nodes and links in a common knowledge graph). For example, a transaction indicating a change in ownership of entity 652 may be stored on the blockchain and sent to a plurality of people to enable, for example, role-based access control, role-based permissions for remote control, identity-based event reporting, etc. May be used by application 630. In an embodiment, the converged process may include shared process flows across applications 630, including a subset of the larger flows involved in one or more of the set of applications 614. For example, a risk management or inspection flow for entity 652 may serve inventory management application 832, asset management application 814, demand management application 824, and supply chain application 812, among others. You can.

In embodiments, robotic process automation 1442 may be provided for a wide range of value chain network processes referred to throughout this disclosure and documents incorporated by reference herein, including without limitation all applications 630. An environment for the development of robotic process automation for a value chain network may include a set of interfaces for developers, where the developer can access selected data sources of the VCN data storage layer 624 and event data 1034, state data ( 1140) or other value chain network data objects 1004 may take inputs from the monitoring system layer 614 and use them as input for classification or prediction, or as inputs to the platform 102, value chain network entity 652, or application 630. ), etc., the artificial intelligence system 1160 can be configured to feed, for example, a neural network. The RPA development environment 1442 may be configured to take output and results 1040 from various applications 630 to again facilitate automated learning and improvement, such as classification, prediction, etc., involved in the steps of the process intended to be automated. You can. In an embodiment, the development environment, and resulting robotic process automation 1442, can be implemented as a software program (e.g., by an operator interacting with various software interfaces of applications 630 involving value chain network entities 652). A combination of both interaction observations 1500 and physical process interaction observations 1510 (e.g., by observing workers interacting with or using machines, equipment, tools, etc. in value chain network 668). It may involve monitoring. In embodiments, observations of software interactions 1500 include interactions between software components with other software components, such as how one application 630 interacts with another application 630 via an API. can do. In embodiments, observations of physical process interactions 1510 may determine how human workers interact with value chain entities 652 (e.g., the location of workers (where a given type of worker is located during a given set of events, process, etc.)). including the path taken through a location), how an operator uses various tools, equipment, and physical interfaces to manipulate a piece of equipment, cargo, container, package, product 650 or other item, various events (e.g., alarms) and responses to alerts); procedures by which workers perform scheduled delivery, movement, maintenance, updates, repair and service processes; procedures by which workers tune or adjust items involved in their workflow. etc.) may include observation (e.g., by a video camera, motion detector, or other sensor, as well as detection of the position, movement, etc. of hardware, such as robotic hardware). Physical process observation 1510 includes the operator's position, angle, force, speed, acceleration, pressure, and torque as the operator moves on hardware, such as on a container or package, or on a piece of equipment involved in handling product with a tool. This may include tracking, etc. These observations include video data, data detected within the machine (such as the positions of elements of the machine detected and reported by position detectors), and the physical characteristics of the human operator's interactions with hardware items (for the purpose of developing training datasets). It can be obtained by any combination of data collected by a wearable device (such as an exoskeleton containing a position detector, force detector, torque detector, etc.) configured to detect. By collecting both software interaction observations 1500 and physical process interaction observations 1510, RPA system 1442 can engage value chain entities 652, e.g., by using software automation in combination with physical robots. The process can be automated more comprehensively.

In an embodiment, robotic process automation 1442 is configured to train a set of physical robots with hardware elements that facilitate performing tasks typically performed by humans. These include walking (including going up and down stairs to deliver packages), climbing (such as climbing a ladder in a warehouse to reach a shelf where product 650 is stored), moving around the facility, Attaching to an item, grasping an item (e.g., using a robotic arm, hand, pincer, etc.), lifting an item, carrying an item, removing and replacing an item, using a tool, etc. It may include a robot that performs tasks.

VALUE CHAIN MANAGEMENT PLATFORM - UNIFIED ROBOTIC PROCESS AUTOMATION FOR DEMAND MANAGEMENT AND SUPPLY CHAIN

In embodiments, methods, systems, components and other elements for an information technology system are provided in this application, including a micro-service architecture, a set of interfaces 702, a set of network connectivity facilities 642, and an adaptive intelligence facility. (614), a data storage facility (624), a data collection system (640), and a cloud-based management VCNP (604) having a monitoring facility (614) coordinated for monitoring and management of a set of value chain network entities (652). ; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; and an integrated set of robotic process automation systems 1442 that provide coordinated automation between various applications 630, including demand management applications, supply chain applications, intelligent product applications, and enterprise resource planning applications for categories of goods. It can be included.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; and a robotic process automation system that provides coordinated automation between at least two types of applications: a set of demand management applications for a category of goods, a set of supply chain applications, a set of intelligent product applications, and a set of enterprise resource management applications. Includes an integrated set.

VALUE CHAIN MANAGEMENT PLATFORM - ROBOTIC PROCESS AUTOMATION SERVICES IN MICROSERVICES ARCHITECTURE FOR VALUE CHAIN NETWORK

In embodiments, the present application provides methods, systems, components and other elements for an information technology system, which includes a micro-services architecture, interfaces 702, coordinated for monitoring and management of a set of value chain network entities 652; ), a cloud-based management VCNP 102 having a set of network connectivity facilities 642, an adaptive intelligence facility 614, a data storage facility 624, a data collection system 640, and a monitoring facility 614. ; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; It may include a set of microservice layers, including an application layer supporting at least one supply chain application and at least one demand management application, wherein the microservice layer provides a set of actions for at least a subset of the applications 630. and a robotic process automation layer 1442 that uses the information collected by the data collection layer 640 and the set of results and activities 1040 accompanying the application of the application layer 630 to automate.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; and a set of microservice layers, including an application layer supporting at least one supply chain application and at least one demand management application, wherein the microservice layer is configured to automate a set of actions for at least a subset of the applications. It includes a robotic process automation layer that uses the information collected by the data collection layer and a set of results and activities followed by the application of the application layer.

VALUE CHAIN MANAGEMENT PLATFORM - ROBOTIC PROCESS AUTOMATION FOR VALUE CHAIN NETWORK PROCESSES

In embodiments, methods, systems, components and other elements for an information technology system are provided in this application, including a micro-service architecture, a set of interfaces 702, a set of network connectivity facilities 642, and an adaptive intelligence facility. (614), a data storage facility (624), a data collection system (640), and a cloud-based management VCNP (102) having a monitoring facility (614) coordinated for monitoring and management of a set of value chain network entities (652). ; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; It may include a set of robotic process automation systems 1442 for automating a set of processes in a value chain network, where robotic process automation systems 1442 may include software used to monitor and manage value chain network entities 652. For a training data set involving a set of user interactions with a set of interfaces 702 of the system, as well as various processes and application outputs and results 1040 that may occur with or within VCNP 102 ) to learn from.

In an embodiment, value chain network entities 652 may include, for example, products, suppliers, producers, manufacturers, retailers, operators, owners, operators, operating facilities, customers, consumers, workers, mobile devices, wearable devices, distributors, etc. Contractors, resellers, supply chain infrastructure facilities, supply chain processes, logistics processes, reverse logistics processes, demand forecasting processes, demand management processes, demand aggregation processes, machines, ships, barges, warehouses, marine ports, airports, shipping routes, Waterways, roads, railways, bridges, tunnels, online retailers, e-commerce sites, demand factors, supply factors, delivery systems, floating assets, points of origin, points of destination, points of storage, points of use, networks, information technology systems, software platforms. , distribution centers, fulfillment centers, containers, container handling facilities, customs, export control, border control, drones, robots, autonomous vehicles, transportation facilities, drones/robots/AV, waterways, port infrastructure facilities, etc. You can.

In embodiments, the robotic process automation layer may perform, for example, but not limited to, selection of quantity of product for an order, selection of a carrier for delivery, selection of a vendor for a component, selection of a vendor for a finished product order, marketing, etc. Selection of product variants, selection of assortment of goods on the shelf, determination of price for finished products, composition of product-related service offers, composition of product bundles, composition of product kits, composition of product packages, product Composition of display, composition of product image, composition of product description, composition of website navigation path related to the product, determination of inventory level for the product, selection of logistics type, composition of schedule for product delivery, composition of logistics schedule, Organizing a set of inputs for machine learning, preparing product documentation, preparing mandatory disclosures regarding the product, organizing the product against a set of local requirements, organizing a set of products for compatibility, organizing a request for proposals, ware Order of equipment to the house, order of equipment to the fulfillment center, classification of product defects in the image, inspection of the product in the image, inspection of product quality data from a set of sensors, a set of on-board diagnostics for the product. examination of data, examination of diagnostic data from Internet of Things systems, examination of sensor data from environmental sensors in a set of supply chain environments, selection of inputs to the digital twin, selection of outputs from the digital twin, from the digital twin Selection of visual elements for the presentation of, Diagnosis of sources of delay in the supply chain, Diagnosis of sources of scarcity in the supply chain, Diagnosis of sources of congestion in the supply chain, Diagnosis of sources of cost overruns in the supply chain , automates processes that may include diagnosing the source of product defects in the supply chain, predicting maintenance requirements in the supply chain infrastructure, etc.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; It may include a set of robotic process automation systems for automating a set of processes in a value chain network, wherein the robotic process automation system may include a set of interfaces to a set of software systems used to monitor and manage value chain network entities. It learns based on a training data set that involves a set of user interactions.

In an embodiment, one of the processes automated by robotic process automation as described in any of the embodiments disclosed herein may involve: In embodiments, RPA involves selection of quantities of products for an order. In an embodiment, one of the processes automated by robotic process automation involves selection of a carrier for a shipment. In embodiments, one of the processes automated by robotic process automation involves selection of vendors for components. In an embodiment, one of the processes automated by robotic process automation involves selection of a vendor for ordering finished products. In an embodiment, one of the processes automated by robotic process automation involves selection of variants of a product for marketing. In an embodiment, one of the processes automated by robotic process automation involves the selection of an assortment of merchandise on a shelf. In an embodiment, one of the processes automated by robotic process automation involves determining a price for a finished product. In an embodiment, one of the processes automated by robotic process automation involves the composition of service offerings related to products. In an embodiment, one of the processes automated by robotic process automation involves the construction of product bundles. In an embodiment, one of the processes automated by robotic process automation involves the construction of product kits. In an embodiment, one of the processes automated by robotic process automation involves the construction of product packages. In an embodiment, one of the processes automated by robotic process automation involves the construction of a product display. In an embodiment, one of the processes automated by robotic process automation involves the construction of product images. In an embodiment, one of the processes automated by robotic process automation involves the construction of product descriptions. In an embodiment, one of the processes automated by robotic process automation involves the construction of website navigation paths related to products. In an embodiment, one of the processes automated by robotic process automation involves determining inventory levels for products. In an embodiment, one of the processes automated by robotic process automation involves selection of a logistics type. In an embodiment, one of the processes automated by robotic process automation involves the construction of schedules for product delivery. In an embodiment, one of the processes automated by robotic process automation involves the construction of a logistics schedule. In an embodiment, one of the processes automated by robotic process automation involves constructing a set of inputs for machine learning. In an embodiment, one of the processes automated by robotic process automation involves preparation of product documentation. In an embodiment, one of the processes automated by robotic process automation involves the preparation of mandatory disclosures regarding products. In an embodiment, one of the processes automated by robotic process automation involves configuration of a product to a set of local requirements. In an embodiment, one of the processes automated by robotic process automation involves organizing sets of products for compatibility. In an embodiment, one of the processes automated by robotic process automation involves the formulation of requests for proposals.

In an embodiment, one of the processes automated by robotic process automation involves ordering equipment for a warehouse. In an embodiment, one of the processes automated by robotic process automation involves ordering equipment for a fulfillment center. In an embodiment, one of the processes automated by robotic process automation involves classification of product defects in images. In an embodiment, one of the processes automated by robotic process automation involves inspection of a product in an image.

In an embodiment, one of the processes automated by robotic process automation involves inspection of product quality data from a set of sensors. In an embodiment, one of the processes automated by robotic process automation involves examination of data from a set of onboard diagnostics for a product. In an embodiment, one of the processes automated by robotic process automation involves examination of diagnostic data from an Internet of Things system. In an embodiment, one of the processes automated by robotic process automation involves reviewing sensor data from environmental sensors within a set of supply chain environments.

In an embodiment, one of the processes automated by robotic process automation involves selection of inputs to a digital twin. In an embodiment, one of the processes automated by robotic process automation involves selection of outputs from a digital twin. In an embodiment, one of the processes automated by robotic process automation involves selection of visual elements for presentation in the digital twin. In an embodiment, one of the processes automated by robotic process automation involves diagnosing sources of delays in the supply chain. In an embodiment, one of the processes automated by robotic process automation involves diagnosis of sources of scarcity in the supply chain. In an embodiment, one of the processes automated by robotic process automation involves diagnosing sources of congestion in the supply chain.

In an embodiment, one of the processes automated by robotic process automation involves diagnosing sources of cost overruns in the supply chain. In an embodiment, one of the processes automated by robotic process automation involves diagnosing the source of product defects in the supply chain. In an embodiment, one of the processes automated by robotic process automation involves predicting maintenance requirements in the supply chain infrastructure.

In embodiments, a set of demand management applications, supply chain applications, intelligent product applications, and enterprise resource management applications may include, for example, supply chain, asset management, risk management, inventory management, demand management, demand forecasting, demand aggregation, and pricing. , location, deployment, promotion, blockchain, smart contracts, infrastructure management, facilities management, analytics, finance, transactions, tax, regulation, identity management, commerce, e-commerce, payments, security, safety, vendor management, process management. , compatibility testing, compatibility management, infrastructure testing, incident management, predictive maintenance, logistics, monitoring, remote control, automation, self-organization, self-healing, self-organization, logistics, reverse logistics, waste reduction, augmented reality, Virtual reality, mixed reality, demand customer profiling, entity profiling, enterprise profiling, worker profiling, workforce profiling, component supply policy management, product design, product configuration, product updates, product maintenance, product support, product testing , warehousing, distribution, fulfillment, kit configuration, kit deployment, kit support, kit updates, kit maintenance, kit modifications, kit management, delivery fleet management, vehicle fleet management, workforce management, maritime fleet management, navigation, routing, It may involve delivery management, opportunity matching, search, advertising, entity discovery, entity discovery, distribution, delivery, enterprise resource planning, etc.

Opportunity Miners for Automated Improvement of Adaptive Intelligence

26, artificial intelligence 1160, automation (robotic process automation 1442), e.g., for one or more of the systems, subsystems, components, applications, etc. of or interacting with VCNP 102 Provided as part of the adaptive intelligence layer 614 is a set of opportunity miners 1460 that can be configured to discover and recommend opportunities to improve one or more elements of the platform 604, including through the addition of It can be. In embodiments, opportunity miner 1460 may be configured or used by developers of AI or RPA solutions to find opportunities for better solutions and optimize existing solutions in value chain network 668. In embodiments, opportunity miner 1460 may comprise a set of systems that collect information within VCNP 102 and within, about, and about a set of value chain network entities 652 and environments. Information collected herein may be used for increased automation and/or intelligence regarding value chain network 668, application 630, value chain network entity 652, or VCNP 102 itself. It has the potential to help identify and prioritize opportunities. For example, opportunity miner 1460 may identify labor-intensive areas and processes in a set of value chain network 668 environments, e.g., by time, type, using cameras, wearables, or other sensors. , and may include a system that observes clusters of value chain network workers by location. These can be used to show places with high labor activity, for example in a ranked or prioritized list, or in visualizations (for example, heat maps showing the dwell time of customers, workers or other individuals on a map of the environment or customers within the environment). Or it can be presented in a heat map showing the path the worker took. In embodiments, analytics 838 may be used to identify which environments or activities would most benefit from automation for the purposes of improved delivery times, alleviation of congestion, and other performance improvements.

In embodiments, opportunity mining may include facilities for requesting appropriate training data sets that may be used to facilitate process automation. Certain types of input, if available, will provide very high value for automation, for example, video data sets that capture very experienced and/or highly specialized workers performing complex tasks. Opportunity miner 1460 may search such video data sets as described herein; However, in the absence of success (or to supplement available data), the platform allows users, such as developers, to collect, for example, software interaction data (e.g., from experts working with programs to perform specific tasks), video data (such as a video showing a set of professionals performing a delivery process, packing process, picking process, container moving process, etc.), and/or physical process observation data (such as video, sensor data, etc.) of the desired type. It may include a system that can specify data. The resulting library of interactions captured in response to the specifications is stored as a data set in the data storage layer 624, for example, for consumption by various applications 630, adaptive intelligence system 614, and other processes and systems. can be captured. In embodiments, the library provides automation maps that can follow instructions, e.g., in a video, such as providing a sequence of steps according to a procedure or protocol, decomposing a procedure or protocol into sub-steps that are candidates for automation, etc. In order to facilitate the development of a video, it may include a video that is specifically developed as an educational video. In embodiments, such videos may be used to automatically develop sequences of labeled instructions that developers can use, for example, to facilitate maps, graphs, or other models of the process to aid in the development of automation for the process; It can be processed by natural language processing. In embodiments, a designated set of training data sets may be configured to act as input to learning. In such cases, the training data can be time-synchronized with other data within the platform 604, such as outputs and results from applications 630, outputs and results from value chain entities 652, etc., such that a given video of a process corresponds to those outputs. and may be associated with outcomes, thereby enabling feedback on learning responsive to outcomes resulting from a given process captured (e.g., on video, or through observation of software interaction or physical process interaction). For example, this may relate to an instructional video, such as a video of a person who may be building or rebuilding (eg, rebuilding a bearing set). This instruction video may include individual steps for reconstruction that may allow staging of the training, for example, to provide instructions for parsing the video into a stage that mimics an expert performing staging on the video. For example, this may include tagging the video to include references to each stage and status (e.g., stage 1 complete, stage 2, etc.). An example of this type could be using artificial intelligence, which can understand that there may be a series of sub-functions that are added to the final function.

In embodiments, opportunity miner 1460 may include methods, systems, processes, components, services, and other elements for mining for opportunities for defining, forming, configuring, and executing smart contracts. handled by data handling layer 608, stored by data storage layer 624, collected by monitoring layer 614 and collection system 640, collected around or from entities 652, or Data collected within platform 604, such as any data obtained from external sources, may be used to recognize beneficial opportunities for the application or composition of smart contracts. For example, pricing information about an entity 652 handled by a pricing application 842, or otherwise collected, may be used to determine whether the same item or items may be priced heterogeneously (in a spot market, futures market, etc.). Opportunity Miner 1460 contracts to buy in one environment at a price below a given threshold and sell in another environment at a price above a given threshold, or vice versa. Likewise, it can provide alerts that indicate opportunities for smart contract formation.

In some examples, as shown in FIG. 26 , adaptive intelligence system 614 may include value converter 1470 . A value translator 1470 may be associated with the demand side of the transaction. Specifically, for example, value converter 1470 can understand the negative currencies of two markets and convert value currencies to other currencies (e.g., as well as fiat currencies that already have explicit conversion capabilities). there is. In some examples, value converter 1470 may be associated with a point in a point-based system (e.g., in a cost-based routing system). In an example embodiment, value converter 1470 may be offered loyalty points that may be convertible to airline seats and/or may be converted to a refund policy for staying in a hotel room. In some examples, different types of entities may be connected as having native price or cost functions that do not always use the same currency or any currency. In another example, value converter 1470 may be used with cost-based routing, where the point system in such cost-based routing systems is not monetary-based, but rather network prioritization or de-prioritization.

BROAD MANAGEMENT PLATFORM

28, additional details of an embodiment of platform 604 are provided, particularly regarding the overall architecture for platform 604. These include, for a cloud-based management platform 604 using a micro-services architecture, a set of network connectivity facilities 642 (which may include or connect to a set of interfaces 702 at various layers of the platform 604); It may include a set of adaptive intelligence facilities or adaptive intelligence systems 614, a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

Accordingly, this application provides methods, systems, components and other elements for an information technology system, including a cloud-based management platform with a micro-services architecture; A set of interfaces, network connection facilities, adaptive intelligence facilities, data storage facilities, and monitoring facilities; and a set of applications to enable an enterprise to manage a set of value chain network entities from the point of origin to the point of customer use.

Also provided in this application are methods, systems, components and other elements for an information technology system, which may include a cloud-based management platform with a micro-services architecture, where the platform can access and configure features of the platform. A set of interfaces for; a set of network connectivity facilities to enable a set of value chain network entities to connect to the platform; A set of adaptive intelligence facilities to automate a set of platform capabilities; A set of data storage facilities for storing data collected and handled by the Platform; It has a set of monitoring facilities for monitoring value chain network entities; The platform hosts a set of applications to enable enterprises to manage a set of value chain network entities from the point of origin of the enterprise's products to the point of customer use.

BROAD MANAGEMENT PLATFORM - DETAILS

29, additional details of an embodiment of platform 604 are provided, particularly regarding the overall architecture for platform 604. These include, for a cloud-based management platform 604 using a micro-services architecture, a set of network connectivity facilities 642 (which may include or connect to a set of interfaces 702 at various layers of the platform 604); It may include a set of adaptive intelligence facilities or adaptive intelligence systems 614, a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

In an embodiment, set of interfaces 702 may include demand management interface 1402 and supply chain management interface 1404.

In an embodiment, a set of network connectivity facilities 642 to enable a set of value chain network entities 652 to connect to a platform 604 may be a 5G network, such as deployed in a supply chain infrastructure facility operated by an enterprise. May include system 1410.

In embodiments, a set of network connectivity facilities 642 to enable a set of value chain network entities 652 to connect to platform 604 may be a network system and/or a cloud computing environment (e.g., data collection system 640). Internet of Things systems (1172), such as those deployed in a supply chain infrastructure facility operated by an enterprise, on or near a value chain network entity (652) when configured to collect and organize this IoT data. ) may include.

In an embodiment, a set of network connectivity facilities 642 to enable a set of value chain network entities 652 to connect to VCNP 102 may be a cognitive networking system deployed in a supply chain infrastructure facility operated by the enterprise ( 1420).

In an embodiment, a set of network connectivity facilities 642 to enable a set of value chain network entities 652 to connect to the VCNP 102 may be connected to a peer-to-peer network, such as those deployed in a supply chain infrastructure facility operated by the enterprise. May include a two-peer network system 1430.

In embodiments, an adaptive intelligence facility or set of adaptive intelligence systems 614 for automating the set of capabilities of platform 604 may be an edge intelligence system, such as deployed in a supply chain infrastructure facility operated by an enterprise ( 1420).

In embodiments, an adaptive intelligence facility or set of adaptive intelligence systems 614 for automating a set of capabilities of platform 604 may include robotic process automation system 1442.

In embodiments, an adaptive intelligence facility or set of adaptive intelligence systems 614 to automate a set of capabilities of a platform 604 deployed in a supply chain infrastructure facility operated by an enterprise, deployed in a network. and/or may include or be integrated with a self-organizing data collection system 1440, such as deployed in a cloud computing environment. This may include elements of the data collection system 640 of the data handling layer 608 that interact or integrate with elements of the adaptive intelligence system 614.

In embodiments, a set of adaptive intelligence facilities or adaptive intelligence systems 614 to automate a set of capabilities of platform 604 that represent the attributes of a set of value chain network entities, e.g., controlled by an enterprise. It may include the same digital twin system 1700.

In embodiments, an adaptive intelligence facility or set of adaptive intelligence systems 614 to automate a set of capabilities of platform 604 such as state data, event data, or other data handled by data handling layer 608. Based on, for example, the value chain may include a smart contract system 848 for automating a set of transactions or interactions between a set of network entities 652.

In embodiments, a data storage facility or set of data storage systems 624 for storing data collected and handled by platform 604 uses distributed data architecture 1122.

In embodiments, blockchain 844 is a set of data storage facilities for storing data collected and handled by the platform.

In embodiments, distributed ledger 1452 is a set of data storage facilities for storing data collected and handled by the platform.

In an embodiment, a set of data storage facilities for storing data collected and handled by the platform uses a graph database 1124, which represents a set of hierarchical relationships of value chain network entities.

In an embodiment, the set of monitoring facilities 614 for monitoring value chain network entities 652 may include, for example, an Internet of Things monitoring system 1172 for collecting data from IoT systems and devices deployed throughout the value chain network. Includes.

In an embodiment, a set of monitoring facilities 614 for monitoring value chain network entity 652 may be installed at or near, e.g., in a value chain environment or within value chain network entity 652, e.g., product 1510. and a set of sensor systems 1462 disposed within or on.

In an embodiment, the set of applications 614 may be of various types, for example, among the sets of supply chain management applications 21004, demand management applications 1502, intelligent product applications 1510, and enterprise resource management applications 1520. Contains a set of applications that may include .

In an embodiment, the set of applications includes asset management application 1530.

In embodiments, value chain network entities 652, as referenced throughout this disclosure, may include, for example, and without limitation, products, suppliers, producers, manufacturers, retailers, operators, owners, operators, operating facilities, customers, Consumers, workers, mobile devices, wearable devices, distributors, resellers, supply chain infrastructure facilities, supply chain processes, logistics processes, reverse logistics processes, demand forecasting processes, demand management processes, demand aggregation processes, machines, ships, barges. , warehouses, maritime ports, airports, sea routes, waterways, roads, railways, bridges, tunnels, online retailers, e-commerce sites, demand factors, supply factors, delivery systems, floating assets, origin points, destination points, storage points, Point of use, networks, information technology systems, software platforms, distribution centers, fulfillment centers, containers, container handling facilities, customs, export control, border control, drones, robots, autonomous vehicles, transport facilities, drones/robots/AV , waterways, port infrastructure facilities, etc.

In an embodiment, platform 604 manages a set of demand factors 1540, a set of supply factors 1550, and a set of value chain infrastructure facilities 1560.

In embodiments, supply factors 1550 as referenced throughout this disclosure may include, for example and without limitation, component availability, material availability, component location, material location, component price, material price, taxation, tariff, Import duties, duties, import controls, export controls, border controls, trade controls, customs, navigation, traffic, congestion, vehicle capacity, vessel capacity, container capacity, package capacity, vehicle availability, vessel availability, container availability, package availability, vehicle location , vessel location, container location, port location, port availability, port capacity, storage availability, storage capacity, warehouse availability, warehouse capacity, fulfillment center location, fulfillment center availability, fulfillment center capacity, asset owner identity, system It may involve compatibility, worker availability, worker capabilities, worker location, commodity price, fuel price, energy price, route availability, route distance, route cost, route safety, etc.

In embodiments, demand factors 1540 as referenced throughout this disclosure may include, for example and without limitation, product availability, product price, delivery timing, refill need, replacement need, manufacturer recall, upgrade need, maintenance. Maintenance Needs, Update Needs, Repair Needs, Supplies Needs, Tastes, Preferences, Inferred Needs, Inferred Desires, Group Needs, Personal Needs, Family Needs, Business Needs, Workflow Needs, Process Needs, Procedure Needs, Processing Needs, Improvement Necessity, diagnostic need, compatibility with systems, compatibility with products, compatibility with styles, compatibility with brands, demographic, psychographic, geolocation, indoor location, destination, route, home location, visiting location, work location, Business location, personality, mood, emotions, customer behavior, business type, business activity, personal activity, wealth, income, purchase history, shopping history, search history, engagement history, clickstream history, website history, online navigation history, groups It may include those involving behavior, family behavior, family membership, customer identity, group identity, business identity, customer profile, business profile, group profile, family profile, declared interest, inferred interest, etc.

In embodiments, supply chain infrastructure facility 1560, as referenced throughout this disclosure, may include, for example and without limitation, ships, container ships, boats, barges, marine ports, cranes, containers, container handling, Shipyards, marine docks, warehouses, distribution, fulfillment, refueling, refueling, nuclear refueling, waste removal, food supply, beverage supply, drones, robots, autonomous vehicles, aircraft, cars, trucks, trains, lifts, forklifts, transport facilities. , conveyors, loading docks, waterways, bridges, tunnels, airports, depots, vehicle stations, train stations, weighing stations, inspection, roads, railways, highways, customs, border control, and other facilities.

In embodiments, the set of applications 614 as referenced throughout this disclosure may include, for example and without limitation, supply chain, asset management, risk management, inventory management, demand management, demand forecasting, demand aggregation, Pricing, positioning, deployment, promotion, blockchain, smart contracts, infrastructure management, facilities management, analytics, finance, transactions, tax, regulation, identity management, commerce, e-commerce, payments, security, safety, vendor management, process. Management, compatibility testing, compatibility management, infrastructure testing, incident management, predictive maintenance, logistics, monitoring, remote control, automation, self-organizing, self-healing, self-organizing, logistics, reverse logistics, waste reduction, augmented reality , virtual reality, mixed reality, demand customer profiling, entity profiling, enterprise profiling, worker profiling, workforce profiling, component supply policy management, product design, product configuration, product updates, product maintenance, product support, product Testing, warehousing, distribution, fulfillment, kit configuration, kit deployment, kit support, kit updates, kit maintenance, kit modifications, kit management, delivery fleet management, vehicle fleet management, workforce management, maritime fleet management, navigation, routing , delivery management, opportunity matching, search, advertising, entity discovery, entity discovery, distribution, delivery, enterprise resource planning, and other applications.

CONTROL TOWER

30, one embodiment of platform 604 is provided. Platform 604 may include various data handling layers 608, a set of network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604), adaptive intelligence facilities, or A micro-service architecture may be used having a set of adaptive intelligence systems (614), a set of data storage facilities or systems (624), and a set of monitoring facilities or systems (808). Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

In an embodiment, the platform 604 includes a supply chain management application 21004, a demand management application 1502, an intelligent product application 1510, which monitors and/or manages a value chain network and a set of value chain network entities 652. ) and demand management information for a category of goods, such as displaying status information, event information, activity information, analytics, reporting, or other elements related to or generated by the set of enterprise resource planning applications 1520, and It may include a user interface 1570 that provides a set of integrated views of a set of supply chain information. Accordingly, the integrated view interface 1570 may, in embodiments, provide the supply chain infrastructure facility 1560 and other value chain components involved as a product 1510 moves from its point of origin through distribution and retail channels to a customer-usable environment. May provide a control tower for an enterprise, spanning a range of assets, such as network entities 652. These may include views of demand factors 1540 and supply factors 1550, allowing the user to develop insight into the connections between factors and control one or both of them with coordinated intelligence. The population of the set of integrated views may be used to determine, for example, which view of the interface 1570 provides the most impactful insights, control features, etc., for the results 1040 or other actions of the adaptive intelligence system 614. By learning, it can adapt over time.

In an embodiment, the user interface includes a voice action assistant 1580.

In an embodiment, the user interface includes a set of digital twins 1700 for presenting a visual representation of a set of attributes of a set of value chain network entities 652.

In embodiments, user interface 1570 may be configured to configure an adaptive intelligent system, such as allowing user selection of properties, parameters, data sources, input for training, feedback for training, views, formats, arrangements, or other elements. (614) Alternatively, it may include the ability to configure adaptive intelligence facilities.

VALUE CHAIN MANAGEMENT PLATFORM - CONTROL TOWER UI FOR DEMAND MANAGEMENT AND SUPPLY CHAIN

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; and a user interface that provides a set of integrated views of a set of demand management information and supply chain information for a category of products.

UNIFIED DATABASE

31, an embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems (614), a set of data storage facilities or systems (624), and a set of monitoring facilities or systems (808). Platform 604 includes a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

In an embodiment, the platform 604 includes a supply chain management application 21004, a demand management application 1502, an intelligent product application 1510, which monitors and/or manages a value chain network and a set of value chain network entities 652. ) and an integrated database 1590 that supports a set of multiple types of applications, such as a set of enterprise resource management applications 1520. Accordingly, the integrated database 1590 may, in embodiments, include supply chain infrastructure facilities 1560 and other value chain networks involved as products 1510 move from the point of origin through distribution and retail channels to customer-use environments. Entities 652, such as 652, may provide integration of data storage, access, and handling for an enterprise across a range of assets. This integration can provide a number of benefits, including reduced need for data entry, consistency across applications 630, reduced latency (and better real-time reporting), reduced need for data transformation and integration, etc. there is. These may include data regarding demand factors 1540 and supply factors 1550 so that applications 630 benefit from information collected, processed, or generated by other applications 630 on platform 604. This allows users to develop insights into the connections between factors and control one or both of them with coordinated intelligence. The population of integrated databases 1590 can determine, for example, which elements of the database 1590 should be made available to which applications, which data structures provide the most benefit, and which data should be stored or cached for immediate retrieval. to determine what data should be stored, what data can be stored versus discarded, what data is most beneficial in supporting the adaptive intelligence system 614, and other uses, e.g., the results of the adaptive intelligence system 614 ( 1040) or can be adapted over time by learning about other movements.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; and an integrated database supporting a set of at least two types of applications among a set of demand management applications for a category of product, a set of supply chain applications, a set of intelligent product applications, and a set of enterprise resource management applications.

In an embodiment, the integrated database that supports a set of demand management applications, a set of supply chain applications, a set of intelligent product applications, and a set of enterprise resource management applications for a category of products is a distributed database.

In an embodiment, a unified database supporting a set of demand management applications, a set of supply chain applications, a set of intelligent product applications, and a set of enterprise resource management applications for product categories utilizes a graph database architecture. In an embodiment, the set of demand management applications includes a demand forecasting application. In an embodiment, the set of demand management applications includes a demand aggregation application. In an embodiment, the set of demand management applications includes a demand activation application.

In an embodiment, the set of supply chain management applications includes a vendor search application. In an embodiment, the set of supply chain management applications includes a route configuration application. In an embodiment, the set of supply chain management applications includes a logistics scheduling application.

UNIFIED DATA COLLECTION SYSTEMS

32, an embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems 1160, a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

In an embodiment, the platform 604 includes a supply chain management application 21004, a demand management application 1502, an intelligent product application 1510, which monitors and/or manages a value chain network and a set of value chain network entities 652. ) and an integrated set of data collection and management systems 640, a set of monitoring facilities or systems 808 that support a set of various types of applications 614, including a set of enterprise resource planning applications 1520. may include. Accordingly, the integrated data collection and management system 640 may, in embodiments, provide the supply chain infrastructure facilities 1560 involved as the product 1510 moves from the point of origin through distribution and retail channels to the environment of customer use. and other value chain network entities 652 may provide integration of data monitoring, search, discovery, collection, access and handling for enterprises or other users. This integration can provide a number of benefits, including reduced need for data entry, consistency across applications 630, reduced latency (and better real-time reporting), reduced need for data transformation and integration, etc. there is. These may include the collection of data regarding demand factors 1540 and supply factors 1550 so that application 630 can access information collected, processed, or generated by other applications 630 on platform 604. Benefits can be gained, allowing users to develop insights into the connections between factors and control one or both of them with coordinated intelligence. Integrated data collection and management system 640 can determine, for example, which elements of data collection and management system 640 should be made available to which applications 630, which data types or sources provide the most benefit; What data should be stored or cached for immediate retrieval, what data can be stored versus discarded, what data is most beneficial in supporting adaptive intelligence system 614, and what data should be made available for other uses. The results 1040 of the adaptive intelligence system 614 may be adapted over time, for example, by learning about other operations. In an example embodiment, integrated data collection and management system 640 may use a unified data schema associated with data collection and management for various applications. This can be a single, tightly bound, point-of-truth database or a set of distributed data systems that can follow a schema that can be common enough to allow a wide variety of applications to consume the same data as received. . For example, sensor data can be pulled from smart products and consumed by logistics applications, finance applications, demand forecasting applications, genetic programming artificial intelligence (AI) applications, etc. to change products. All of these applications can consume data from the data framework. In one example, this could come from a distributed ledger or blockchain that could contain transaction data for purchases and sales, or an indication of whether an event has occurred. In some example embodiments, as data moves through the supply chain, these data flows may occur through distributed databases, relational databases, any type of graph database, etc., which may be part of an integrated data collection and management system 640. You can. In another example, integrated data collection and management system 640 may utilize memory, which may be dedicated memory on assets, in tags or as part of the memory structure of the device itself, which can be accessed from robust pipelines tied to value chain network entities. It can originate from In another example, integrated data collection and management system 640 may use classic data integration capabilities, which may include adapting protocols such that the protocols ultimately arrive at an integrated system or schema.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; and an integrated set of data collection systems that support a set of at least two types of applications among a set of demand management applications, a set of supply chain applications, a set of intelligent product applications, and a set of enterprise resource management applications for a category of goods. It can be included.

In an embodiment, the integrated set of data collection systems includes a set of crowdsourcing data collection systems. In an embodiment, the integrated set of data collection systems includes a set of Internet of Things data collection systems. In an embodiment, the integrated set of data collection systems includes a set of self-organizing sensor systems. In embodiments, the integrated set of data collection systems includes a set of data collection systems that interact with a networked product.

In an embodiment, the integrated set of data collection systems includes a set of mobile data collectors deployed in a set of value chain network environments operated by an enterprise. In an embodiment, the integrated set of data collection systems includes a set of edge intelligence systems deployed in a set of value chain network environments operated by an enterprise. In an embodiment, the integrated set of data collection systems includes a set of crowdsourcing data collection systems. In an embodiment, the integrated set of data collection systems includes a set of Internet of Things data collection systems. In an embodiment, the integrated set of data collection systems includes a set of self-organizing sensor systems. In embodiments, the integrated set of data collection systems includes a set of data collection systems that interact with a networked product. In an embodiment, the integrated set of data collection systems includes a set of mobile data collectors deployed in a set of value chain network environments operated by an enterprise. In an embodiment, the integrated set of data collection systems includes a set of edge intelligence systems deployed in a set of value chain network environments operated by an enterprise.

Unified IoT Monitoring Systems

33, one embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems 1160, a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

In an embodiment, the platform 604 includes a supply chain management application 21004, a demand management application 1502, an intelligent product application ( of an Internet of Things system 1172 that provides coordinated monitoring of various value chain entities 652 in the service of a set of multiple applications 630 of various types, such as a set of enterprise resource management applications 1520) May contain integrated sets.

Accordingly, the integrated set of Internet of Things systems 1172 may, in embodiments, provide supply chain infrastructure facilities 1560 that accompany products 1510 as they move from the point of origin through distribution and retail channels to customer-use environments. and other value chain network entities 652. Such integration can provide a number of benefits, including reduced need for data input, consistency across applications 630, reduced latency, real-time reporting and recognition, reduced need for data transformation and integration, etc. These may include an Internet of Things system 1172 used in connection with demand factors 1540 and supply factors 1550, such that an application 630 can be used to provide an Internet of Things system ( 1172), users can benefit from information collected, processed, or generated by an integrated set of there is. The integrated set of Internet of Things systems 1172 may determine, for example, which elements of the integrated set of Internet of Things systems 1172 should be made available to which applications 630 and which IoT systems 1172 would benefit the most. What data should be stored or cached for immediate retrieval, what data can be stored versus discarded, what data is most beneficial in supporting adaptive intelligence system 614, and utilized for other purposes. Adaptations may be made over time, such as by learning about the results 1040 or other actions of the adaptive intelligence system 614 to determine whether they should be enabled. In some examples, an integrated set of Internet of Things (IoT) system 1172 may be IoT devices that may be installed in various environments. One goal of the integrated set of Internet of Things systems 1172 involves city-wide deployment, where collectively a set of IoT devices can be connected by wide area network protocols (e.g., longer range protocols). This may be city or town-wide coordination. In another example, an integrated set of Internet of Things systems 1172 may involve connecting a mesh of devices across several different distribution facilities. IoT devices can identify aggregates for each warehouse, and warehouses can communicate with each other using IoT devices. IoT devices can be configured to process data without using the cloud.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications integrated with the Platform to enable enterprise users of the Platform to manage a set of value chain network entities from point of origin to point of customer use; and an Internet of Things system that provides coordinated monitoring of a set of at least two types of applications among a set of demand management applications, a set of supply chain applications, a set of intelligent product applications, and a set of enterprise resource management applications for a category of goods. May contain integrated sets.

In an embodiment, an integrated set of Internet of Things systems includes a set of smart home Internet of Things devices to enable monitoring of a set of demand factors and a set of supply chain infrastructure facilities to enable monitoring of a set of supply factors. It includes a set of Internet of Things devices placed in close proximity.

In an embodiment, an integrated set of Internet of Things systems is provided to enable monitoring of a set of supply factors and a set of workplace Internet of Things devices to enable monitoring of a set of demand factors for a set of business customers and a supply chain infrastructure to enable monitoring of a set of supply factors. It includes a set of Internet of Things devices deployed in close proximity to a set of facilities.

In embodiments, an integrated set of Internet of Things systems enables monitoring of a set of supply factors and a set of Internet of Things devices to monitor a set of consumer goods stores to enable monitoring of a set of demand factors for a set of consumers. It includes a set of Internet of Things devices deployed in close proximity to a set of supply chain infrastructure facilities.

In embodiments, Internet of Things systems as referenced throughout this disclosure may include, for example and without limitation, camera systems, lighting systems, motion detection systems, weighing systems, inspection systems, machine vision systems, environmental sensor systems, etc. , on-board sensor systems, on-board diagnostic systems, environmental control systems, sensor-enabled network switching and routing systems, RF sensing systems, magnetic sensing systems, pressure monitoring systems, vibration monitoring systems, temperature monitoring systems, heat flow monitoring systems, biological measurements. systems, chemical measurement systems, ultrasonic monitoring systems, radiography systems, LIDAR-based monitoring systems, access control systems, penetrating wave detection systems, SONAR-based monitoring systems, radar-based monitoring systems, computed tomography systems, magnetic resonance imaging systems, network monitoring. It may include systems, etc.

Machine Vision Feeding Digital Twin

34, an embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems 1160, a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

In embodiments, platform 604 may include a machine vision system 1600 and a digital twin system 1700, where machine vision system 1600 is an adaptive intelligence system (including artificial intelligence 1160). A digital twin system that can be enabled by a set of 614) and can be used as an interface or component of interface 702 such that an operator can monitor twins 1700 of various value chain network entities 652. Data is supplied to (1700). Machine vision system 1600 and digital twin system 1700 may include a supply chain management application 21004, a demand management application 1502, which monitors and/or manages a value chain network and a set of value chain network entities 652; It can operate cooperatively on a set of multiple applications 630 of various types, such as a set of intelligent product applications 1510 and enterprise resource management applications 1520.

Accordingly, machine vision system 1600 and digital twin system 1700 may, in embodiments, include image-based monitoring (using automated processing of image data), a wide range of facilities, devices (using automated processing of image data), Systems, environments, and assets, such as supply chain infrastructure facilities 1560 and other value chain network entities 652 involved as products 1510 move from their point of origin through distribution and retail channels to environments where they are consumed by customers. In addition, it is possible to provide not only the representation of the image in the digital twin 1700 but also data extracted from the image. This integration can provide a number of benefits, including improved monitoring, improved visualization and insights, improved visibility, and more. These may include machine vision systems 1600 and digital twin systems 1700 used in connection with demand factors 1540 and supply factors 1550 so that applications 630 can be used in conjunction with other applications 630 on platform 604. ) can benefit from information collected, processed, or generated by the machine vision system 1600 and the digital twin system 1700, allowing users to develop insights about connections between factors and create coordinated intelligence. You can control one or both of these. Machine vision system 1600 and/or digital twin system 1700 may utilize any of the elements collected and/or processed by machine vision system 1600 and/or digital twin system 1700 for any application 630, for example. What data should be enabled, what elements and/or content will provide the greatest benefit, what data should be stored or cached for immediate retrieval, what data can be stored versus discarded, and what data should be used by adaptive intelligence systems ( 614) over time, e.g., by learning about the results 1040 or other actions of the adaptive intelligent system 614 to determine whether it is most beneficial to support the system and whether it should be made available for other uses. It can be adapted.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; It includes a set of at least two types of applications among a set of supply chain applications, a set of demand management applications, a set of intelligent product applications, and a set of enterprise resource management applications, and may have a machine vision system and a digital twin system, and the machine The vision system supplies data to the digital twin system.

In embodiments, the set of supply chain applications and demand management applications are any of those described throughout this disclosure or in documents incorporated by reference herein.

In embodiments, a set of supply chain applications and demand management applications may include, for example, and without limitation, inventory management, demand forecasting, demand aggregation, pricing, blockchain, smart contracts, positioning, placement, promotions, analytics, financial , trading, arbitrage, customer identity management, store planning, shelf planning, customer path planning, customer path analysis, commerce, e-commerce, payments, customer relationship management, sales, marketing, advertising, bidding, customer monitoring, customer process monitoring, Customer relationship monitoring, collaborative filtering, customer profiling, customer feedback, similarity analysis, customer clustering, product clustering, seasonality factor analysis, customer behavior tracking, customer behavior analysis, product design, product configuration, A/B testing, product fluctuation analysis, Augmented Reality, Virtual Reality, Mixed Reality, Customer Demand Profiling, Customer Mood, Emotion or Affect Detection, Customer Mood, Emotion of Impact Analysis, Business Entity Profiling, Customer Enterprise Profiling, Demand Matching, Location-Based Targeting, Location- Base offer, point-of-sale interface, point-of-use interface, search, advertising, entity discovery, entity discovery, enterprise resource planning, workforce management, customer digital twin, product pricing, product bundling, product and service bundling, segmentation by product type, upsell offer configuration , one or more involving customer feedback engagement, customer surveys, etc.

In embodiments, a set of supply chain applications and demand management applications may include, but are not limited to, supply chain, asset management, risk management, inventory management, blockchain, smart contracts, infrastructure management, facilities management, analytics, finance, trading, and tax. , regulation, identity management, commerce, e-commerce, payments, security, safety, vendor management, process management, compatibility testing, compatibility management, infrastructure testing, incident management, predictive maintenance, logistics, monitoring, remote control, automation, self-service. -Organization, self-healing, self-organizing, logistics, reverse logistics, waste reduction, augmented reality, virtual reality, mixed reality, supply chain digital twin, vendor profiling, supplier profiling, manufacturer profiling, logistics entity profiling, Enterprise profiling, worker profiling, workforce profiling, component supply policy management, warehousing, distribution, fulfillment, delivery fleet management, vehicle fleet management, workforce management, maritime fleet management, navigation, routing, delivery management, opportunity matching, It may include one or more of the following: search, entity discovery, entity discovery, distribution, delivery, enterprise resource planning, or other applications.

In embodiments, a set of supply chain applications and demand management applications may include asset management, risk management, inventory management, blockchain, smart contracts, analytics, finance, trading, tax, regulatory, identity management, commerce, e-commerce, payments, and security. , safety, compatibility testing, compatibility management, incident management, predictive maintenance, monitoring, remote control, automation, self-organization, self-healing, self-organization, waste reduction, augmented reality, virtual reality, mixed reality, product design, Product configuration, product updates, product maintenance, product support, product testing, kit configuration, kit deployment, kit support, kit update, kit maintenance, kit modification, kit management, product digital twin, opportunity matching, search, advertising, entities One or more of the following applications: discovery, entity search, permutation, simulation, user interface, application programming interface, connectivity management, natural language interface, voice/conversation interface, robot interface, touch interface, haptic interface, vision system interface, enterprise resource planning, or other applications. may include, without limitation.

In embodiments, a set of supply chain applications and demand management applications may include, without limitation, operations, finance, asset management, supply chain management, demand management, human resources management, product management, risk management, regulatory and compliance management, inventory management, and infrastructure. Structural management, facilities management, analytics, transactions, taxes, identity management, vendor management, process management, project management, operations management, customer relationship management, workforce management, incident management, research and development, sales management, marketing management, fleet management, It may include one or more of opportunity analysis, decision support, strategic planning, forecasting, resource management, asset management, or other applications.

In an embodiment, the machine vision system includes an artificial intelligence system that is trained to recognize types of value chain assets based on a labeled dataset of images of these types of value chain assets.

In embodiments, a digital twin provides an indication of the type of asset based on the output of an artificial intelligence system.

In an embodiment, the machine vision system includes an artificial intelligence system that is trained to recognize types of activities involving a set of value chain entities based on a labeled data set of images of these types of activities.

In embodiments, the digital twin presents an indicator of the type of activity based on the output of the artificial intelligence system.

In embodiments, the machine vision system comprises an artificial intelligence system that is trained to recognize safety risks accompanying value chain entities based on a training data set that includes a set of images of value chain network activity and a set of value chain network safety results. Includes.

In embodiments, a digital twin presents indicators of risk based on the output of an artificial intelligence system.

In an embodiment, the machine vision system includes an artificial intelligence system that is trained to predict delays based on a training data set that includes a set of images of value chain network activity and a set of value chain network timing results.

In embodiments, the digital twin provides an indication of potential delays based on the output of an artificial intelligence system.

As referenced elsewhere in this application and in documents incorporated by reference, with respect to value chain network entities 652 and related processes and applications (such as any of the technologies or systems described throughout this disclosure) Artificial intelligence may be used, inter alia, to facilitate: (a) optimization, automation and/or control of various functions, workflows, applications, features, resource utilization and other factors, (b) various states, entities. , recognition or diagnosis of patterns, events, situations, behaviors, or other factors; and/or (c) anticipation of various conditions, events, circumstances or other factors. As artificial intelligence improves, a large array of domain-specific and/or general artificial intelligence systems have become available and are likely to continue to proliferate. As developers seek solutions to domain-specific problems, such as those related to the value chain entities 652 and applications 630 described throughout this disclosure, they may need to consider selecting an artificial intelligence model (e.g., what set of Neural networks, machine learning systems, expert systems, etc. face challenges in discovering and choosing which inputs to choose) and which inputs can enable effective and efficient use of artificial intelligence for a given problem. As mentioned above, Opportunity Miner 1460 can help discover opportunities for increased automation and intelligence; However, once the opportunity is discovered, the selection and configuration of artificial intelligence solutions still presents significant challenges, which are likely to continue to grow as artificial intelligence solutions proliferate.

One set of solutions to this challenge is an artificial intelligence repository 3504 configured to enable the collection, organization, recommendation, and presentation of a related set of artificial intelligence systems based on one or more properties of the domain and/or domain-related problems. am. In embodiments, artificial intelligence repository 3504 may include downloads of related artificial intelligence applications, links to artificial intelligence systems, or other connections (e.g., via APIs, ports, connectors, or other interfaces) to cloud-deployed artificial intelligence systems. may include a set of interfaces to an artificial intelligence system, such as enabling the establishment of Artificial intelligence repository 3504 may provide domain-specific input, data or other entities and/or for solving specific types of problems (e.g., prediction, NLP, image recognition, pattern recognition, motion detection, path optimization, etc.). It may include descriptive content for each of the various artificial intelligence systems, such as metadata or other descriptive material indicating the suitability of the system to operate on the system. In embodiments, artificial intelligence repository 3504 may be organized by categories such as domain, input type, processing type, output type, computational requirements and capabilities, cost, energy usage, and other factors. In embodiments, the interface to application repository 3504 may take input from a developer and/or from a platform (e.g., from opportunity miner 1460) indicating one or more properties of a problem that can be addressed through artificial intelligence; , may provide a set of recommendations for subsets of artificial intelligence solutions that may represent favorable candidates based on the developer's domain-specific problems, such as through an artificial intelligence attribute search engine. Search results or recommendations may, in embodiments, be made by, for example, asking the developer to indicate or select elements of a favorable model, but also by, for example, a similarity matrix, k-means clustering, or similar developers, similar domain-specific problems, and/ or by clustering using other clustering techniques to associate similar artificial intelligence solutions, based at least in part on collaborative filtering. Artificial intelligence repository 3504 may operate using smart contracts and/or blockchain features to automate purchases, licensing, payment tracking, settlement of transactions, or other features, as well as links to ratings, reviews, and related content. , and mechanisms for provisioning, licensing, delivery, and payment (including allocation of payments to affiliates and/or contributors).

Referring to FIG. 43, artificial intelligence system 1160 performs data processing, data analysis, simulation generation, and analysis, simulation, decision making, and prediction related to data processing of one or more of the value chain entities 652. A machine learning model (3000) to perform can be defined. Machine learning model 3000 is an algorithm and/or statistical model that performs a specific task by relying instead on patterns and inferences without using explicit instructions. Machine learning model 3000 is not explicitly programmed to perform a specific task but rather builds one or more mathematical models based on training data to make predictions and/or decisions. Machine learning model 3000 may receive input of sensor data including event data 1034 and state data 1140 associated with one or more of the value chain entities 652 as training data. Sensor data input to the machine learning model 3000 is used to perform analysis, simulation, decision making, and prediction decisions related to data processing, data analysis, simulation generation, and simulation analysis of one or more of the value chain entities 652. It can be used to train a machine learning model 3000 to perform. Machine learning model 3000 may also use input data from a user or users of an information technology system. Machine learning model 3000 may include an artificial neural network, decision tree, support vector machine, Bayesian network, genetic algorithm, any other suitable form of machine learning model, or a combination thereof. The machine learning model 3000 is configured to learn through supervised learning, unsupervised learning, reinforcement learning, self-learning, feature learning, sparse dictionary learning, anomaly detection, association rules, combinations thereof, or any other suitable algorithm for learning. It can be configured.

Artificial intelligence system 1160 may also define a digital twin system 1700 to create a digital replica of one or more of the value chain entities 652. The digital replicas of one or more of the value chain entities 652 may use real-time sensor data to provide a real-time virtual representation of the value chain entities 652 and one or more possible future states of the one or more value chain entities 652 . Provides a simulation of The digital replica exists concurrently with one or more value chain entities 652 being replicated. The digital replica may, in embodiments, represent one or more simulations of both the physical elements and characteristics of the one or more value chain entities 652 being replicated and their dynamics throughout the lifestyle of the one or more value chain entities 652 being replicated. to provide. The digital replica may be used, for example, during the design phase before one or more value chain entities are constructed or manufactured, or during or after the construction or manufacture of one or more value chain entities, during high stresses, when component wear may be an issue. After a period of time, during peak throughput operation, after one or more hypothetical or planned improvements have been made to one or more value chain entities 652, or during any other suitable hypothetical situation, one or more value chain entities (652) A virtual simulation of one or more value chain entities 652 may be provided by allowing virtual extrapolation of sensor data to simulate the state of 652). In some embodiments, machine learning model 3000 can be used to predict when one or more components of one or more value chain entities 652 may fail, such as predicting possible improvements to one or more value chain entities 652. predicting and/or suggesting possible improvements to one or more value chain entities 652, such as changes to timing settings, arrangements, components, or any other suitable changes to the value chain entities 652. By doing so, virtual situations for simulation with digital replicas can be automatically predicted. The digital replica provides a simulation of one or more value chain entities 652 during both the design and operation phases of one or more value chain entities 652, as well as a simulation of virtual operating conditions and configurations of one or more value chain entities 652. Allowed. The digital replica can be used to measure temperature, wear, light, and vibration within, on, and around each component of one or more value chain entities 652, as well as, in some embodiments, within one or more value chain entities 652. It allows for valuable analysis and simulation of one or more value chain entities by facilitating the observation and measurement of almost any type of metric, including: In some embodiments, machine learning model 3000 may process sensor data, including event data 1034 and state data 1140, to define simulation data for use by digital twin system 1700. . The machine learning model 3000, for example, receives state data 1140 and event data 1034 related to a specific value chain entity 652 among the plurality of value chain entities 652, and receives state data 1140 and a series of operations on the state data 1140 and the event data 1034 to format the event data 1034 into a format suitable for use by the digital twin system 1700 in creating a digital replica of the value chain entity 652. can be performed. For example, one or more value chain entities 652 may include robots configured to augment products on an adjacent assembly line. Machine learning model 3000 may collect data from one or more sensors located on the robot, near the robot, within the robot, and/or around the robot. The machine learning model 3000 may perform operations on sensor data, process the sensor data into simulation data, and output the simulation data to the digital twin system 1700. Digital twin simulation 1700 may use simulation data to create one or more digital replicas of a robot, where the simulation includes metrics including, for example, temperature, wear, speed, rotation, and vibration of the robot and its components. do. The simulation may be a substantially real-time simulation that allows a human user of the information technology to view a simulation of the robot, its associated metrics, and metrics associated with its components in substantial real-time. A simulation may be a predictive or virtual situation that allows a human user of information technology to view a predictive or virtual simulation of a robot, metrics related to it, and metrics related to its components.

In some embodiments, machine learning model 3000 and digital twin system 1700 process sensor data and create a digital replica of a set of value chain entities 652 to identify related values of the value chain entities. May facilitate group design, real-time simulation, predictive simulation, and/or virtual simulation. The digital replica of the set of value chain entities may use substantially real-time sensor data to provide a substantially real-time virtual representation of the set of value chain entities and to provide a simulation of one or more possible future states of the set of value chain entities. A digital replica exists concurrently with the set of value chain entities being replicated. The digital replica provides one or more simulations of both the physical elements and characteristics of the set of replicated value chain entities and their dynamics, in embodiments spanning the lifestyle of the set of replicated value chain entities. One or more simulations include a visual simulation, such as a wire-frame virtual representation of one or more value chain entities 652, viewable on a monitor using an augmented reality (AR) device, or using a virtual reality (VR) device. can do. The visual simulation may be manipulated by a human user of the information technology system, such as zooming or highlighting components of the simulation and/or providing an exploded view of one or more value chain entities 652. The digital replica may be used in such a way that component wear is a problem, for example, during high stress, for example, during the design phase before one or more value chain entities are constructed or manufactured, or during or after the construction or manufacturing of one or more value chain entities. After a period of time has elapsed, during peak throughput operation, after one or more hypothetical or planned improvements have been made to the set of value chain entities, or during any other suitable hypothetical situation, the status of the set of value chain entities It can provide a virtual simulation of a set of value chain entities by allowing virtual extrapolation of sensor data to simulate. In some embodiments, machine learning model 3000 can, for example, predict possible improvements to a set of value chain entities, predict when one or more components of the set of value chain entities may fail, and/ or by suggesting possible improvements to the set of value chain entities, such as timing settings, arrangements, changes to components, or any other suitable changes to the value chain entities 652. Hypothetical situations can be automatically predicted. The digital replica allows simulation of the set of value chain entities during both the design and operation phases of the set of value chain entities as well as the simulation of the virtual operating conditions and configuration of the set of value chain entities. The digital replica may be a digital replica of virtually any type of lithography, including temperature, wear, light, vibration, etc., within, on, and around each component of the set of value chain entities, as well as, in some embodiments, within the set of value chain entities. By facilitating the observation and measurement of metrics, it allows valuable analysis and simulation of one or more value chain entities. In some embodiments, machine learning model 3000 may process sensor data, including event data 1034 and state data 1140, to define simulation data for use by digital twin system 1700. . The machine learning model 3000, for example, receives state data 1140 and event data 1034 related to a specific value chain entity 652 among the plurality of value chain entities 652, and receives state data 1140 and performing a series of operations on the state data 1140 and event data 1034 to format the event data 1034 into a format suitable for use by the digital twin system 1700 in creating a digital replica of the set of value chain entities. It can be done. For example, a set of value chain entities may include a die machine configured to place a product on a conveyor belt, a conveyor belt configured to place the product on a conveyor belt, and a conveyor belt configured to add parts to the product as it moves along the assembly line. It may include a plurality of robots. Machine learning model 3000 may collect data from one or more sensors located on, near, within, and/or about the die machine, conveyor belt, and plurality of robots, respectively. The machine learning model 3000 may perform operations on sensor data, process the sensor data into simulation data, and output the simulation data to the digital twin system 1700. Digital twin simulation 1700 may use simulation data to create one or more digital replicas of a die machine, a conveyor belt, and a plurality of robots, where the simulation can be performed, for example, by a die machine, a conveyor belt, and a plurality of robots, and Includes metrics including temperature, wear, speed, rotation, and vibration of that component. The simulation may be a substantially real-time simulation that allows a human user of the information technology to view a simulation of a die machine, a conveyor belt, and a plurality of robots, their associated metrics, and metrics associated with their components in substantially real-time. A simulation may be a predictive or virtual situation that allows a human user of the information technology to view a predictive or virtual simulation of a die machine, conveyor belt, and plurality of robots, metrics related thereto, and metrics related to their components.

In some embodiments, machine learning model 3000 may prioritize the collection of sensor data for use in simulations of digital replicas of one or more of the value chain entities 652. Machine learning model 3000 can be trained using sensor data and user input, thereby learning which types of sensor data are most effective for creating a digital replica simulation of one or more of the value chain entities 652. For example, machine learning model 3000 may discover that a particular value chain entity 652 has dynamic characteristics such as component wear and throughput that are affected by temperature, humidity, and load. The machine learning model 3000 can prioritize the collection of sensor data related to temperature, humidity, and load through machine learning, and output the prioritized types of sensor data to the digital twin system 1700. You can prioritize processing with simulation data for . In some embodiments, machine learning model 3000 can be used to simulate the value chain entity 652 so that more and/or better data of prioritized types can be used in the simulation of value chain entities 652 via its digital replica. It may be suggested to users of information technology systems that more and/or different sensors of prioritized types be implemented in the information technology and value chain systems near and around entity 652.

In some embodiments, machine learning model 3000 determines which types of sensor data to process as simulation data for transmission to digital twin system 1700 based on one or both of modeling goals and the quality or type of sensor data. It can be configured to learn to determine what should be done. The modeling goal may be a goal set by a user of the information technology system or may be predicted or learned by the machine learning model 3000. Examples of modeling goals may include, for example, collection, simulation, and modeling of heat, power, component wear, and other metrics of conveyor belts, assembly machines, one or more products, and other components of the value chain. It involves creating a digital replica that can show the throughput dynamics on the assembly line. Machine learning model 3000 may be configured to learn to determine the types of sensor data that need to be processed into simulation data for transmission to digital twin system 1700 to achieve such a model. In some embodiments, machine learning model 3000 may analyze what type of sensor data is being collected, the quality and quantity of the sensor data being collected, and what the sensor data being collected represents, and which type. Determine, predict, analyze, and/or determine whether sensor data of the sensor data is relevant and/or not relevant to achieving modeling objectives, and is used by digital twin system 1700 in achieving modeling objectives. Determine, predict, analyze, and/or determine to prioritize, improve, and/or achieve the quality and quantity of sensor data processed into simulation data.

In some embodiments, a user of the information technology system may input modeling goals into machine learning model 3000. Machine learning model 3000 may determine whether one or more types of sensors located within, on, or near a value chain entity or a plurality of value chain entities relevant to achieving the modeling goal are sufficient to achieve the modeling goal and/or can learn to analyze training data to output to users of the information technology system suggestions as to which types of sensor data are most relevant to achieving modeling goals, such as which are not sufficient, e.g., adding sensors; By removing, or repositioning, different configurations of sensor types, learning how different configurations of these types of sensors can better facilitate the achievement of modeling goals by the machine learning model 3000 and digital twin system 1700. can do. In some embodiments, machine learning model 3000 automatically increases or increases the collection rate, processing, storage, sampling rate, bandwidth allocation, bitrate, and other properties of sensor data collection to achieve or better achieve modeling goals. can be reduced. In some embodiments, machine learning model 3000 may increase the collection rate, processing, storage, sampling rate, bandwidth allocation, bitrate, and other properties of sensor data collection to achieve or better achieve modeling goals. Suggestions or predictions regarding reductions may be made to users of the information technology system. In some embodiments, the machine learning model 3000 may use sensor data, simulation data, transfer, Current, and/or future digital replica simulations may be used. In some embodiments, modeling goals automatically generated by machine learning model 3000 may be automatically implemented by machine learning model 3000. In some embodiments, modeling goals automatically generated by machine learning model 3000 may be suggested to a user of the information technology system and may be proposed by the user, such as after modifications to the modeling goals suggested by the user have been made. It may be implemented only after acceptance and/or partial acceptance.

In some embodiments, a user may input one or more modeling objectives, for example, by entering one or more modeling commands into an information technology system. One or more modeling commands may be used to generate a digital replica simulation of a value chain entity 652 or a set of multiple 652 value chain entities, e.g., the machine learning model 3000 and the digital twin system 1700. It may include a command for, and may include a command for a digital replica simulation, which is one or more of real-time simulation and virtual simulation. The modeling commands may also include, for example, parameters for what type of sensor data should be used, a sampling rate for the sensor data, and other parameters for the sensor data to be used in one or more digital replica simulations. In some embodiments, machine learning model 3000 may be configured to predict modeling commands, such as by using previous modeling commands as training data. Machine learning model 3000 facilitates simulation of one or more of value chain entities 652, which may be useful, for example, in the management of value chain entities 652 and/or allows a user to interact with value chain entities 652. Predicted modeling commands may be suggested to users of the information technology system to facilitate identification of potential problems or possible improvements thereto.

In some embodiments, machine learning model 3000 may be configured to evaluate a set of virtual simulations of one or more value chain entities 652 . The set of virtual simulations may be generated by a machine learning model 3000, as a result of one or more modeling commands, by one or more modeling objectives, by a prediction by the machine learning model 3000, as a result of one or more modeling commands, or by a combination thereof. and may be generated by the digital twin system 1700. Machine learning model 3000 may evaluate a set of virtual simulations based on one or more metrics defined by a user, one or more metrics defined by machine learning model 3000, or a combination thereof. In some embodiments, machine learning model 3000 may evaluate each virtual simulation of a set of virtual simulations independently of one another. In some embodiments, machine learning model 3000 relates one or more of the virtual simulations of the set of virtual simulations to each other, for example, by ranking the virtual simulations or creating a hierarchy of virtual simulations based on one or more metrics. This can be evaluated.

In some embodiments, machine learning model 3000 is one or more models to facilitate human understanding of the output of machine learning model 3000, as well as information and insights related to the cognition and processes of machine learning model 3000. It may include an interpretability system, i.e., one or more model interpretability systems that determine not only “what” the machine learning model 3000 is outputting, but also “why” the machine learning model 3000 outputs that output. allows human understanding of what is being done and what processes are causing the machine learning model 3000 to produce its output. One or more model interpretability systems may also be used to improve and guide the training of machine learning model 3000, to help debug machine learning model 3000, and to help recognize bias in machine learning model 3000. Can be used by human users. One or more model interpretability systems include linear regression, logistic regression, generalized linear models (GLM), generalized additive models (GAM), decision trees, decision rules, RuleFit, Naive Bayes classifier, K-nearest neighbor algorithm, partial Dependency plots, Individual Conditional Expectations (ICE), Accumulated Local Effects (ALE) plots, feature interactions, permuted feature importance, global proxy models, local proxy (LIME) models, bounds rules i.e. anchors, Shapley values, SHAP (Shapley additive explanations), feature visualization, network dissection, or any other suitable machine learning interpretability implementation. In some embodiments, one or more model interpretability systems may include a model data set visualization system. The model data set visualization system is configured to automatically provide human users of the information technology system with visual analysis related to the distribution of the values of the sensor data, simulation data, and data nodes of the machine learning model 3000.

In some embodiments, the machine learning model 3000 may include and/or implement an embedded model interpretability system, such as a Bayesian case model (BCM) or a glass box. The Bayesian case model uses Bayesian case-based reasoning, prototypic classification, and clustering to facilitate human understanding of data such as sensor data, simulation data, and data nodes in the machine learning model 3000. In some embodiments, the model interpretability system includes a glass box interpretability method, such as a Gaussian process, to facilitate human understanding of data such as sensor data, simulation data, and data nodes of the machine learning model 3000. and/or may be implemented.

In some embodiments, machine learning model 3000 may include and/or implement testing using concept activation vectors (TCAVs). TCAV allows a machine learning model 3000 to traditionally be “executed,” “non-executed,” or “powered” by a process that includes defining a concept, determining concept activation vectors, and calculating directional derivatives. )", "not powered", "robot", "human", "truck", or "ship". By learning human-interpretable concepts, objects, states, etc., TCAV allows the machine learning model 3000 to generate useful information related to the value chain entity 652 and data collected therefrom that are easily understood by human users of the information technology system. You can allow output in any format.

In some embodiments, machine learning model 3000 may be an artificial neural network, e.g., a connectionist system, configured to “learn” to perform a task by considering examples and without being explicitly programmed with task-specific rules, and/ It may exist or include this. Machine learning model 3000 may be based on a collection of connected units and/or nodes that may act like artificial neurons that may emulate in some way the neurons of a biological brain. Units and/or nodes may each have one or more connections to other units and/or nodes. The unit and/or node transmits information, e.g., one or more signals, to another unit and/or node, processes signals received from the other unit and/or node, and sends the processed signal to the other unit and/or node. It can be configured to forward to . One or more of the units and/or nodes and connections therebetween may have one or more numerical “weights” assigned to them. The assigned weights may be configured to facilitate learning, or training, of the machine learning model 3000. The assigned weights may increase and/or decrease one or more signals between one or more units and/or nodes, and in some embodiments may have one or more thresholds associated with one or more of the weights. One or more thresholds may be configured such that signals are only transmitted between one or more units and/or nodes if the signal and/or aggregate signal intersects the threshold. In some embodiments, units and/or nodes may be assigned to multiple layers, each layer having one or both inputs and outputs. The first layer may be configured to receive training data, transform at least a portion of the training data, and transmit signals related to the training data and the transformation to the second layer. The final layer may be configured to output an estimate, conclusion, product, or other result of the processing of one or more inputs by the machine learning model 3000. Each layer may perform one or more types of transformation, and one or more signals may pass through one or more of the layers more than once. In some embodiments, machine learning model 3000 may utilize deep learning and is configured at least in part as a deep neural network, a deep trust network, a recurrent neural network, and/or a convolutional neural network, such as by being configured to include one or more hidden layers. It can be modeled and/or configured.

In some embodiments, machine learning model 3000 may be and/or include a decision tree, e.g., a tree-based predictive model configured to identify one or more observations and determine one or more conclusions based on the input. . Observations may be modeled as one or more “branches” of a decision tree, and conclusions may be modeled as one or more “leafs” of the decision tree. In some embodiments, the decision tree may be a classification tree. A classification tree may include one or more leaves representing one or more class labels, and one or more branches representing one or more connections of features that are configured to lead to the class labels. In some embodiments, the decision tree may be a regression tree. A regression tree can be constructed so that one or more target variables can take on continuous values.

In some embodiments, machine learning model 3000 may be and/or include a support vector machine, e.g., a set of related supervised learning methods configured for use in one or both classification and regression-based modeling of data. You can. A support vector machine can be configured to predict whether a new example belongs to one or more categories, and these one or more categories are constructed during training of the support vector machine.

In some embodiments, machine learning model 3000 may be configured to perform regression analysis to determine and/or estimate a relationship between one or more inputs and one or more features of the one or more inputs. Regression analysis may include linear regression, where machine learning model 3000 may compute a single line to best fit the input data according to one or more mathematical criteria.

In embodiments, inputs to machine learning model 3000 (such as a regression model, Bayesian network, supervised model, or other type of model) may be used to test the impact of various inputs on the accuracy of model 3000, for example. This can be tested, for example, by using a set of test data that is independent of the data set used to create and/or train the machine learning model. For example, inputs to a regression model including single inputs, pairs of inputs, triplets, etc. may be removed to determine whether the absence of an input causes a significant deterioration in the success of model 3000. This can help recognize inputs that are actually correlated (e.g., linear combinations of the same basic data), overlapping, etc. Comparison of model success involves choosing between alternative sets of input data that provide similar information, i.e., selecting which input (a few It can help identify (among similar inputs) etc. Accordingly, testing input variation and its impact on model validity can be used to prune or improve model performance for any of the machine learning systems described throughout this disclosure.

In some embodiments, machine learning model 3000 may be and/or include a Bayesian network. A Bayesian network can be a probabilistic graphical model constructed to represent a set of random variables and the conditional independence of the set of random variables. Bayesian networks can be constructed to represent random variables and conditional independence through directed acyclic graphs. Bayesian networks may include one or both dynamic Bayesian networks and influence diagrams.

In some embodiments, machine learning model 3000 may be defined through supervised learning, i.e., one or more algorithms configured to build a mathematical model of a set of training data containing one or more inputs and desired outputs. Training data may consist of a set of training examples, each of which has one or more inputs and a desired output, i.e., a guidance signal. Each training example may be represented in machine learning model 3000 by an array and/or vector, that is, a feature vector. Training data can be expressed in the machine learning model 3000 by a matrix. The machine learning model 3000 may learn one or more functions through iterative optimization of the objective function, thereby learning to predict outputs associated with new inputs. Once optimized, the objective function can provide machine learning model 3000 with the ability to accurately determine outputs for inputs other than those included in the training data. In some embodiments, machine learning model 3000 may be defined through one or more supervised learning algorithms, such as active learning, statistical classification, regression analysis, and similarity learning. Active learning may include interactively querying users and/or information sources to label new data points with desired outputs, by the machine learning model AILD102T. Statistical classification may involve identifying, by the machine learning model 3000, new observations to which a set of subcategories, i.e., subpopulations, belong based on a training data set containing observations with known categories. Regression analysis may include estimating, by machine learning model 3000, a relationship between a dependent variable, i.e., an outcome variable, and one or more independent variables, i.e., predictors, covariates, and/or characteristics. Similarity learning may involve learning, by machine learning model 3000, from examples using a similarity function, which is designed to measure how similar or related two objects are.

In some embodiments, machine learning model 3000 may use unsupervised learning, i.e., one or more algorithms configured to build a mathematical model of a set of data containing only the input by discovering structure in the data, such as grouping or clustering of data points. It can be defined through In some embodiments, machine learning model 3000 may learn from labeled, classified, or uncategorized test data, i.e., training data. An unsupervised learning algorithm may involve learning by a machine learning model 3000 by identifying commonalities in training data and reacting based on the presence or absence of the identified commonalities in new pieces of data. In some embodiments, machine learning model 3000 may generate one or more probability density functions. In some embodiments, machine learning model 3000 generates a set of observations according to one or more pre-specified criteria, e.g., according to a similarity metric wherein internal compactness, separation, estimated density, and/or graph connectivity are factors. It can be learned by performing cluster analysis by assigning to a subset, that is, a cluster.

In some embodiments, machine learning model 3000 may be defined through semi-supervised learning, i.e., one or more algorithms using training data, where some training examples may be missing training labels. Semi-supervised learning can be weakly supervised learning, and the training labels can be noisy, limited, and/or inaccurate. Noisy, limited, and/or inaccurate training labels may be cheaper and/or less labor-intensive to generate, and thus machine learning models 3000 may be used to produce larger signals for less cost and/or labor. Allows training on a set of training data.

In some embodiments, machine learning model 3000 may be trained through reinforcement learning, such as one or more algorithms that use dynamic programming techniques to allow machine learning model 3000 to train by taking actions in the environment to maximize a cumulative reward. can be defined. In some embodiments, training data is expressed as a Markov decision process.

In some embodiments, machine learning model 3000 may be defined through self-learning, and machine learning model 3000 may be provided with training data without external compensation and external teaching, for example, by using crossbar adaptive array (CAA). It is configured to train using. CAA may calculate decisions regarding actions and/or emotions regarding the resulting situation in a crossbar manner, thereby driving the teaching of the machine learning model 3000 by the interaction between cognition and emotion.

In some embodiments, machine learning model 3000 is defined through feature learning, i.e., one or more inputs provided during training, e.g., one or more algorithms designed to discover increasingly accurate and/or appropriate representations of the training data. It can be. Feature learning may include training through principal component analysis and/or cluster analysis. The feature learning algorithm may include attempting to preserve the input training data while also transforming the input training data such that the transformed input training data is useful to the machine learning model 3000. In some embodiments, machine learning model 3000 may be configured to transform input training data before performing one or more classifications and/or predictions on the input training data. Accordingly, machine learning model 3000 may be configured to reconstruct input training data from one or more unknown data generating distributions without necessarily complying with an implausible configuration of the input training data according to the distribution. In some embodiments, the feature learning algorithm may be performed by machine learning model 3000 in a supervised, unsupervised, or semi-supervised manner.

In some embodiments, machine learning model 3000 may be defined through anomaly detection, i.e., by identifying rare and/or outlier instances of one or more items, events, and/or observations. Rare and/or outlier instances can be identified by instances that differ significantly from most patterns and/or properties in the training data. Unsupervised anomaly detection may involve detecting, by the machine learning model 3000, an anomaly in an unlabeled training data set under the assumption that the majority of the training data is “normal.” Supervised anomaly detection may include training on a data set in which at least a portion of the training data is labeled as “normal” and/or “abnormal.”

In some embodiments, machine learning model 3000 may be defined through robot learning. Robotic learning involves the creation of one or more curricula by a machine learning model 3000 - a curriculum is a sequence of learning experiences - and exploration guided by a machine learning model 3000 and humans by a machine learning model 3000. It may involve the cumulative acquisition of new skills through social interaction with others. Acquisition of new skills may be facilitated by one or more guidance mechanisms such as active learning, maturation, motor synergy, and/or imitation.

In some embodiments, machine learning model 3000 may be defined through association rule learning. Association rule learning may involve discovering relationships between variables in a database, by a machine learning model 3000, to identify strong rules using some measure of “interest.” Association rule learning may involve identifying, learning, and/or evolving rules to store, manipulate, and/or apply knowledge. Machine learning model 3000 may be configured to learn by identifying and/or utilizing a set of relationship rules, where relationship rules collectively represent the knowledge captured by machine learning model 3000. Association rule learning may include one or more of a learning classifier system, inductive logic programming, and artificial immune system. A learning classifier system is an algorithm that can combine a discovery component, such as one or more genetic algorithms, with a learning component, such as one or more algorithms for supervised learning, reinforcement learning, or unsupervised learning. Guided logic programming may include rule learning using logic programming to represent, by the machine learning model 3000, one or more of the input examples, background knowledge, and hypotheses determined by the machine learning model 3000 during training. You can. The machine learning model 3000 may be configured to derive a hypothesized logic program that entails all positive examples, given a set of examples represented as a logical database of facts and encodings of known background knowledge.

In embodiments, another set of solutions that may be deployed alone or in conjunction with other elements of the platform, including artificial intelligence repository 3504, such as monitoring system 640 and, in some cases, various value chain entities 652 ) may include a set of functional imaging capabilities 3502, which may include a physical process observation system 1510 and/or a software interaction observation system 1500 for monitoring. Functional imaging system 3502, in embodiments, may provide significant insight into the types of artificial intelligence that are likely to be most effective in solving particular types of problems. As noted elsewhere in this disclosure and in documents incorporated by reference into this application, as computing and networking systems grow in scale, complexity, and interconnection, they become subject to information overload, noise, network congestion, energy waste, etc. indicates a problem. As the Internet of Things grows to hundreds of billions of devices and virtually countless potential interconnections, optimization becomes very difficult. One source for insight is the human brain, which has faced similar challenges and, over thousands of years, has evolved rational solutions to a wide range of very difficult optimization problems. The human brain operates as a large-scale neural network organized into interconnected modular systems, each of which addresses a specific problem, from, among other things, the regulation of biological systems and the maintenance of homeostasis, to the detection of a wide range of static and dynamic patterns, and the recognition of threats and opportunities. There is some degree of adaptation to solve the problem. Functional imaging 3502, such as functional magnetic resonance imaging (fMRI), electroencephalography (EEG), computed tomography (CT), and other brain imaging systems, allows patterns of brain activity to be recognized in real time and other information such as behavior, stimulation, etc. Information has been improved to the point where it can be temporally associated with environmental condition data, gestures, eye movements, and other information, and thus via functional imaging, alone or in combination with other information collected by the monitoring system 808. The platform may determine which brain modules, operations, systems, and/or functions are utilized during the performance of a set of tasks or activities, such as those involving software interaction 1500, physical process observation 1510, or a combination thereof. You can decide and classify. This classification is intended to provide, for example, initial construction of robotic process automation (RPA) systems 1442 that automate tasks performed by expert humans, and/or capabilities similar to the set of modules and functions of the human brain when performing an activity. or assist in the selection and/or construction of a set of artificial intelligence solutions, e.g., from artificial intelligence repository 3504, including a set of functionality. Accordingly, the platform can selectively and automatically configure a set of artificial intelligence capabilities for a robotic process automation system based on matching properties between one or more artificial intelligence systems and one or more biological systems, such as a brain system, and a functional imaging system. It may include a system that takes input from. The selection and configuration of inputs to robotic process automation and/or artificial intelligence that are configured based at least in part on functional imaging of the brain while an operator performs a task, such as visual input in which the brain's visual system is highly activated. Selection of acoustic inputs (e.g., images from a camera), selection of acoustic inputs (e.g., images from a camera) for which the brain's auditory system is highly activated, selection of chemical inputs (e.g., chemical sensors) for which the brain's olfactory system is highly active, etc. there is. Therefore, biologically aware robotic process automation systems can be improved by having an initial configuration or iterative improvement, automatically or by imaging-guided information collected as the operator performs expert tasks that would benefit from automation. Can be guided under developer control.

Referring to FIG. 27 , platform 604 relates to elements of an adaptive intelligence layer 614 that facilitates improved edge intelligence, including, in particular, an adaptive edge computing management system 1400 and an edge intelligence system 1420. ) Additional details of the embodiment are provided. These factors include “edge” computing, storage, and storage by varying data storage and processing locations, for example, between on-device storage, local systems, on the network, and in the cloud (e.g., optimized by AI). and a set of systems that adaptively manage processing. These elements facilitate dynamic definition by a user, such as a developer, operator, or host of platform 102, of what constitutes an “edge” for the purposes of a given application. The remoteness of some environments (such as in geographies with poor cellular network infrastructure), for example, environments where data connections are slow or unreliable (e.g., facilities do not have good access to cellular networks); Shielding or interference (e.g. density of network-enabled systems, thick metal hulls of container ships, thick metal container walls, underwater or underground locations, or presence of large metal objects (e.g. bolts, hulls, containers, etc.) interfere with networking performance edge computing capabilities provide a peer-to-peer network of devices on an environment's local area network, and/or congestion (e.g., when there are many devices seeking access to limited networking facilities). 652, or on the computing capabilities of local value chain entities 652. For example, in an environment with a limited set of computational and/or networking resources, tasks may be defined and deployed to operate in the current situation (e.g. may be intelligently load balanced based on network availability, latency, congestion, etc.), e.g., one type of data is prioritized for processing, one workflow is prioritized over another, etc. When strong data connections are available (e.g., good backhaul facilities exist), edge computing capabilities can improve input/output performance or reduce latency, for example, by caching frequently used data at that location. reduce, etc. Accordingly, adaptive definitions and specifications of where edge computing operations are enabled may be under the control of a developer or operator, or optionally, for example, a financial entity (e.g., for the environment). 652), or automatically determined, for example by an expert system or an automated system, based on detected network conditions for the network as a whole.

In embodiments, edge intelligence 1420 may include adaptation of edge computations (including where computations occur within the various available networking resources, how networking occurs (e.g., by protocol selection), where data storage occurs, etc.) enables the value of edge computation capability (ROI) across more than one application, including any combination and subset of applications 630 described herein or in documents incorporated by reference herein. , yield, and cost information such as failure costs), such as considering QoS, latency requirements, congestion, and costs as understood and prioritized based on awareness of requirements, and prioritization. -Application aware.

35, one embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems 1160, a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

In an embodiment, the platform 604 includes a supply chain management application 1500, a demand management application 1502, an intelligent product application ( 21004) and an integrated set of adaptive edges that provide coordinated edge computation and other edge intelligence 1420 capabilities for a set of multiple applications 630 of various types, such as a set of enterprise resource planning applications 1520. May include computing and other edge intelligence systems 1420. In embodiments, the edge intelligence capabilities of the systems and methods described herein may include on-premises edge devices and resources, such as local area network resources, and network edge devices, such as those deployed at the edge of a cellular network or within a peripheral data center. , but are not limited to, both of these, for example, to perform intelligent processing tasks at these edge locations prior to delivering data or other matter to primary or core cellular network commands or to a central data center. As described in , edge intelligence can be deployed.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; and a set of integrated edge calculations that provide coordinated edge calculations for a set of at least two types of applications among a set of demand management applications, a set of supply chain applications, a set of intelligent product applications, and a set of enterprise resource management applications for a category of goods. May include a set of adaptive edge computing systems.

Accordingly, adaptive edge computing and other edge intelligence systems 1420 may, in embodiments, provide supply chain infrastructure facilities ( 1560) and other value chain network entities 652 may provide intelligence for monitoring, managing, controlling, or otherwise handling a wide range of facilities, devices, systems, environments, and assets. This integration can deliver numerous benefits, including improved monitoring, improved remote control, improved autonomy, improved forecasting, improved classification, improved visualization and insights, improved visibility, and more. These may include adaptive edge computing and other edge intelligence systems 1420 used in connection with demand factors 1540 and supply factors 1550 so that applications 630 can interact with other applications 630 on platform 604. Adaptive edge computing and other edge intelligence systems 1420 may benefit from information collected, processed, or generated by users to develop insights about connections between factors and interact with them with coordinated intelligence. You can control one or both. For example, but not limited to, coordinated intelligence may include, but is not limited to, analysis and processing to monitor data streams, as described herein, for the purposes of classification, prediction, or some other type of analytical modeling. . These coordinated intelligence methods and systems can be applied in an automated manner where different combinations of intelligence assets are applied. As an example, within an industrial environment, a coordinated intelligence system may monitor signals coming from machines deployed in the environment. A coordinated intelligence system may classify, predict, or perform some other intelligent analysis, in combination, for the purpose of determining the state of a machine, for example, a machine in a deteriorated state, a critical state, or some other state. You can. Determination of the condition may cause the control system to change the control regime, for example, to slow down or shut down a machine in a deteriorating condition. In embodiments, a coordinated intelligence system may coordinate across multiple entities, such as a value chain, supply chain, etc. For example, monitoring of deteriorating machines in an industrial environment may occur simultaneously with analyzes related to parts suppliers and availability, product supply and inventory forecasting, or some other coordinated intelligent behavior. Adaptive edge computing and other edge intelligence systems 1420 may determine, for example, which elements collected and/or processed by adaptive edge computing and other edge intelligence systems 1420 should be made available to which applications 630 . , which elements and/or content provide the most benefit, which data should be stored or cached for immediate retrieval, which data can be stored versus discarded, and which data supports the adaptive intelligence system 614 may be adapted over time, e.g., by learning about other actions or results 1040 of other adaptive intelligence systems 614, to determine whether they are most beneficial for use and whether they should be made available for other uses. there is.

36, in embodiments, an integrated set of adaptive edge computing systems that provide coordinated edge computation include a classification system 1610 (e.g., an image classification system, an object type recognition system, etc.), a video processing system (1612) (e.g., video compression systems), signal processing system 1614 (e.g., analog-to-digital conversion system, digital-to-analog conversion system, RF filtering system, analog signal processing system, multiplexing system, statistical signal processing system, signal filtering system, natural language processing system, sound processing system, ultrasonic processing system, etc.), data processing system 1630 (e.g., data filtering system, data integration system, data extraction system, data loading system, data conversion system, point cloud processing systems, data normalization systems, data cleansing systems, data deduplication systems, graph-based data storage systems, object-oriented data storage systems, etc.), prediction systems 1620 (e.g., motion prediction systems, output prediction systems, activity prediction systems, faults prediction system, failure prediction system, accident prediction system, event prediction system, event prediction system, etc.), configuration system 1630 (e.g., protocol selection system, storage configuration system, peer-to-peer network configuration system, power management system, self-organizing system, self-healing system, handshake negotiation system, etc.), artificial intelligence system 1160 (e.g., clustering system, variational system, machine learning system, expert system, rule-based system, deep learning system, etc.), system Management and control systems 1640 (e.g., autonomous control systems, robotic control systems, RF spectrum management systems, network resource management systems, storage management systems, data management systems, etc.), robotic process automation systems, analysis and modeling systems 1650 (e.g., data visualization system, clustering system, similarity analysis system, random forest system, physical modeling system, interaction modeling system, simulation system, etc.), entity discovery system, security system 1670 (e.g., cybersecurity system, biometric recognition system, intrusion detection system, firewall system, etc.), rule engine system, workflow automation system, opportunity discovery system, testing and diagnostic system (1660), software image propagation system, virtualization system, digital twin system, Internet of Things monitoring system, routing It includes a wide range of systems such as systems, switching systems, indoor positioning systems, geolocation systems, etc.

In an embodiment, the interface is a user interface to a command center dashboard through which an enterprise orchestrates a set of value chain entities related to a type of product.

In an embodiment, the interface is a user interface of a local management system located in an environment hosting a set of value chain entities.

In embodiments, a local management system user interface facilitates configuration of a set of network connections to an adaptive edge computing system.

In embodiments, a local management system user interface facilitates configuration of a set of data storage resources for an adaptive edge computing system.

In embodiments, a local management system user interface facilitates configuration of a set of data integration capabilities for an adaptive edge computing system.

In embodiments, a local management system user interface facilitates configuration of a set of machine learning input resources for an adaptive edge computing system.

In embodiments, a local management system user interface facilitates configuration of a set of power resources to support an adaptive edge computing system.

In embodiments, a local management system user interface facilitates configuration of a set of workflows managed by an adaptive edge computing system.

In an embodiment, the interface is a user interface of a mobile computing device with a network connection to an adaptive edge computing system.

In an embodiment, the interface is an application programming interface.

In embodiments, an application programming interface facilitates the exchange of data between an adaptive edge computing system and a cloud-based artificial intelligence system.

In embodiments, an application programming interface facilitates data exchange between an adaptive edge computing system and a real-time operating system of a cloud data management platform.

In embodiments, an application programming interface facilitates data exchange between computational facilities of an adaptive edge computing system and a cloud data management platform.

In embodiments, an application programming interface facilitates the exchange of data between an adaptive edge computing system and a set of environmental sensors that collect data about the environment hosting a set of value chain network entities.

In embodiments, an application programming interface facilitates the exchange of data between an adaptive edge computing system and a set of sensors that collect data about a product.

In embodiments, an application programming interface facilitates the exchange of data between an adaptive edge computing system and a set of sensors that collect data posted by intelligent products.

In embodiments, an application programming interface facilitates the exchange of data between an adaptive edge computing system and a set of sensors that collect data posted by a set of Internet of Things systems deployed in an environment hosting a set of value chain network entities. Let's do it.

In embodiments, a set of demand management applications, supply chain applications, intelligent product applications, and enterprise resource management applications may include, for example, any of the applications mentioned throughout this disclosure or in documents incorporated by reference herein. It can be included.

Unified Adaptive Intelligence

37, one embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems 1160, a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

In an embodiment, VCNP 102 includes a set of demand management applications 1502, a set of supply chain applications 1500, a set of intelligent product applications 1510, a set of enterprise resource planning applications 1520, and assets for categories of goods. It may include an integrated set of adaptive intelligence systems 614 that provide coordinated intelligence for a diverse set of applications, such as a set of management applications 1530.

In embodiments, an integrated set of adaptive intelligence systems may be used to serve various layers of the platform 604, as well as, without limitation, whether used for edge intelligence or for intelligence within the network, within the application, or within the cloud. , edge intelligence system (1420), classification system (1610), data processing system (1612), signal processing system (1614), artificial intelligence system (1160), prediction system (1620), configuration system (1630), control system ( 1640), analysis system 1650, test/diagnostic system 1660, security system 1670, and other systems, including a wide variety of systems described throughout this disclosure and in documents incorporated by reference in this application. do. These include neural networks, deep learning systems, model-based systems, expert systems, machine learning systems, rule-based systems, opportunity miners, robotic process automation systems, data transformation systems, data extraction systems, data loading systems, genetic programming systems, image classification systems, Video compression system, analog-digital conversion system, digital-analog conversion system, signal analysis system, RF filtering system, motion prediction system, object type recognition system, point cloud processing system, analog signal processing system, signal multiplexing system, data fusion system , sensor fusion system, data filtering system, statistical signal processing system, signal filtering system, signal processing system, protocol selection system, storage configuration system, power management system, clustering system, fluctuation system, machine learning system, event prediction system, autonomous control system, robot control system, robotic process automation system, data visualization system, data normalization system, data cleansing system, data deduplication system, graph-based data storage system, intelligent agent system, object-oriented data storage system, self-organizing system, self-organizing system -Healing system, self-organizing system, self-organizing map system, cost-based routing system, handshake negotiation system, entity discovery system, cybersecurity system, biometric recognition system, natural language processing system, dialogue processing system, speech recognition system, sound Processing systems, sonication systems, artificial intelligence systems, rule engine systems, workflow automation systems, opportunity discovery systems, physical modeling systems, test systems, diagnostic systems, software image propagation systems, peer-to-peer network configuration systems, RF spectrum Management system, network resource management system, storage management system, data management system, intrusion detection system, firewall system, virtualization system, digital twin system, Internet of Things monitoring system, routing system, switching system, indoor location system, geolocation system, parsing system, semantic filtering system, machine vision system, fuzzy logic system, recommender system, conversation management system, etc.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; and a set of integrated adaptive intelligence systems that provide coordinated intelligence for a set of demand management applications for a category of goods, a set of supply chain applications, a set of intelligent product applications, and a set of enterprise resource management applications. .

In an embodiment, the integrated set of adaptive intelligence systems includes a set of artificial intelligence systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of neural networks. In an embodiment, the integrated set of adaptive intelligence systems includes a set of deep learning systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of model-based systems.

In an embodiment, the integrated set of adaptive intelligence systems includes a set of expert systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of machine learning systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of rule-based systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of opportunity miners.

In an embodiment, the integrated adaptive intelligence system set includes a robotic process automation system set. In an embodiment, the integrated adaptive intelligence system set includes a data transformation system set. In an embodiment, the integrated adaptive intelligence system set includes a data extraction system set. In an embodiment, the integrated set of adaptive intelligence systems includes a set of data loading systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of genetic programming systems.

In an embodiment, the set of integrated adaptive intelligence systems includes a set of image classification systems. In an embodiment, the integrated adaptive intelligence system set includes a video compression system set. In an embodiment, the integrated adaptive intelligence system set includes a set of analog-to-digital conversion systems. In an embodiment, the integrated adaptive intelligence system set includes a digital-to-analog conversion system set. In an embodiment, the integrated adaptive intelligence system set includes a signal analysis system set.

In an embodiment, the set of integrated adaptive intelligence systems includes a set of RF filtering systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of motion prediction systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of object type recognition systems. In an embodiment, the integrated adaptive intelligence system set includes a point cloud processing system set. In an embodiment, the integrated adaptive intelligence system set includes a set of analog signal processing systems.

In an embodiment, the integrated adaptive intelligence system set includes a signal multiplexing system set. In an embodiment, the integrated adaptive intelligence system set includes a data fusion system set. In an embodiment, the integrated adaptive intelligence system set includes a sensor fusion system set. In an embodiment, the integrated adaptive intelligence system set includes a data filtering system set. In an embodiment, the integrated adaptive intelligence system set includes a statistical signal processing system set.

In an embodiment, the integrated adaptive intelligence system set includes a signal filtering system set. In an embodiment, the integrated adaptive intelligence system set includes a signal processing system set. In an embodiment, the integrated adaptive intelligence system set includes a protocol selection system set. In an embodiment, the integrated set of adaptive intelligence systems includes a set of stored configuration systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of power management systems.

In an embodiment, the integrated adaptive intelligence system set includes a clustering system set. In an embodiment, the integrated set of adaptive intelligent systems includes a set of variable systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of machine learning systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of event prediction systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of autonomous control systems.

In an embodiment, the integrated adaptive intelligence system set includes a robot control system set. In an embodiment, the integrated adaptive intelligence system set includes a robotic process automation system set. In an embodiment, the integrated set of adaptive intelligence systems includes a set of data visualization systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of data normalization systems. In an embodiment, the integrated adaptive intelligence system set includes a data cleansing system set.

In an embodiment, the integrated set of adaptive intelligence systems includes a set of data deduplication systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of graph-based data storage systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of intelligent agent systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of object-oriented data storage systems.

In an embodiment, the integrated adaptive intelligence system set includes a self-organizing system set. In an embodiment, the integrated set of adaptive intelligence systems includes a set of self-healing systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of self-organizing systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of self-organizing map systems.

In an embodiment, the set of integrated adaptive intelligence systems includes a set of cost-based routing systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of handshake negotiation systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of entity discovery systems. In an embodiment, the integrated adaptive intelligence system set includes a cybersecurity system set.

In an embodiment, the integrated set of adaptive intelligence systems includes a set of biometric systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of natural language processing systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of conversation processing systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of speech recognition systems.

In an embodiment, the integrated adaptive intelligence system set includes a sound processing system set. In an embodiment, the integrated adaptive intelligence system set includes a ultrasonic processing system set. In an embodiment, the integrated set of adaptive intelligence systems includes a set of artificial intelligence systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of rule engine systems.

In an embodiment, the integrated adaptive intelligence system set includes a workflow automation system set. In an embodiment, the integrated set of adaptive intelligence systems includes a set of opportunity discovery systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of physical modeling systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of test systems.

In an embodiment, the integrated adaptive intelligence system set includes a diagnostic system set. In an embodiment, the integrated adaptive intelligence system set includes a software image propagation system set. In an embodiment, the set of integrated adaptive intelligence systems includes a set of peer-to-peer network configuration systems. In an embodiment, the set of integrated adaptive intelligence systems includes a set of RF spectrum management systems.

In an embodiment, the integrated adaptive intelligence system set includes a network resource management system set. In an embodiment, the integrated adaptive intelligence system set includes a storage management system set. In an embodiment, the integrated adaptive intelligence system set includes a data management system set. In an embodiment, the integrated set of adaptive intelligence systems includes a set of intrusion detection systems.

In an embodiment, the integrated adaptive intelligence system set includes a firewall system set. In an embodiment, the integrated adaptive intelligence system set includes a virtualized system set. In an embodiment, the integrated adaptive intelligence system set includes a digital twin system set. In an embodiment, the integrated set of adaptive intelligence systems includes a set of Internet of Things monitoring systems.

In an embodiment, the integrated set of adaptive intelligence systems includes a set of routing systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of transition systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of indoor location systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of geolocation systems.

In an embodiment, the integrated set of adaptive intelligence systems includes a set of parsing systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of semantic filtering systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of machine vision systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of fuzzy logic systems.

In an embodiment, the set of integrated adaptive intelligence systems includes a set of recommender systems. In an embodiment, the integrated set of adaptive intelligence systems includes a set of conversation management systems. In an embodiment, the set of interfaces includes a demand management interface and a supply chain management interface. In an embodiment, the interface is a user interface to a command center dashboard through which an enterprise orchestrates a set of value chain entities related to a type of product.

In an embodiment, the interface is a user interface of a local management system located in an environment hosting a set of value chain entities. In embodiments, the local management system user interface facilitates configuration of a set of network connections to the adaptive intelligence system. In embodiments, a local management system user interface facilitates configuration of a set of data storage resources for an adaptive intelligence system. In embodiments, a local management system user interface facilitates configuration of a set of data integration capabilities for an adaptive intelligence system.

In embodiments, a local management system user interface facilitates configuration of a set of machine learning input resources for an adaptive intelligence system. In embodiments, a local management system user interface facilitates configuration of a set of power resources to support an adaptive intelligence system. In embodiments, the local management system user interface facilitates configuration of a set of workflows managed by the adaptive intelligence system.

In an embodiment, the interface is a user interface of a mobile computing device with a network connection to an adaptive intelligence system.

In an embodiment, the interface is an application programming interface. In embodiments, an application programming interface facilitates the exchange of data between an adaptive intelligence system and a cloud-based artificial intelligence system. In embodiments, an application programming interface facilitates data exchange between an adaptive intelligence system and a real-time operating system of a cloud data management platform.

In embodiments, an application programming interface facilitates data exchange between the adaptive intelligence system and the computational facilities of the cloud data management platform.

In embodiments, an application programming interface facilitates the exchange of data between an adaptive intelligence system and a set of environmental sensors that collect data about the environment hosting a set of value chain network entities. In embodiments, an application programming interface facilitates the exchange of data between an adaptive intelligence system and a set of sensors that collect data about a product.

In embodiments, an application programming interface facilitates the exchange of data between an adaptive intelligence system and a set of sensors that collect data posted by an intelligent product.

In embodiments, an application programming interface facilitates the exchange of data between an adaptive intelligence system and a set of sensors that collect data published by a set of Internet of Things systems deployed in an environment hosting a set of value chain network entities. do.

In embodiments, the set of demand management applications, supply chain applications, intelligent product applications, and enterprise resource management applications may include any of the applications mentioned throughout this disclosure or in documents incorporated by reference into this application.

In an embodiment, adaptive intelligence system layer 614 is configured to train and deploy artificial intelligence systems to perform value chain-related tasks. For example, adaptive intelligence systems layer 614 may manage container fleets, design the logistics system, control one or more aspects of the logistics system, select packaging attributes for packages in the value chain, and satisfy regulatory requirements. It may be utilized to design processes to reduce waste generation, automate processes to mitigate waste generation (e.g., solid waste or wastewater), and/or perform other suitable tasks related to the value chain.

In some of these embodiments, one or more digital twins may be utilized by adaptive intelligence system layer 614. A digital twin is a physical object (e.g., asset, device, product, package, container, vehicle, vessel, etc.), environment (e.g., facility), individual (e.g., customer or worker), or other entity (e.g., may refer to a digital representation of a value chain network entity 652 (including any of the value chain network entities 652 described herein), or a combination thereof. Other examples of physical assets include containers (e.g. boxes, shipping containers, crates, pallets, barrels, etc.), goods/products (e.g. widgets, food, household products, toys, clothing, water, gas, oil, etc.) , equipment, etc.), components (e.g., chips, boards, screens, chipsets, wires, cables, cards, memory, software components, firmware, parts, connectors, housing, etc.), furniture (e.g., tables, counters, workstations, stands, etc.). Examples of devices include computers, sensors, vehicles (e.g., cars, trucks, tankers, trains, forklifts, cranes, etc.), equipment, conveyor belts, etc. Examples of environments may include facilities (e.g., factories, refineries, warehouses, retail locations, storage buildings, parking lots, airports, commercial buildings, residential buildings, etc.), roads, waterways, cities, countries, land, etc. . Examples of different types of physical assets, devices, and environments are referenced throughout this disclosure.

In embodiments, a digital twin may include other digital twins (e.g., through reference, or by partial or complete integration). For example, a digital twin of a package may include a digital twin of the container and one or more digital twins of each of one or more products enclosed within the container. Taking this example one step further, one or more digital twins of a package could be included in a digital twin of a vehicle crossing a road, or located on a digital twin of a stall within a digital twin of a warehouse, where the Digital twins will include digital twins of other physical assets and devices.

In embodiments, a digital representation of a digital twin collectively defines a set of characteristics, attributes, and/or parameters of a represented physical asset, device, or environment, possible behavior or activity, and/or possible state or condition, among other things. may contain a set of data structures (e.g., classes of objects). For example, the set of characteristics of a physical asset may include the type of physical asset, the shape and/or dimensions of the asset, the mass of the asset, the density of the asset, the material(s) of the asset, the physical properties of the material(s), and the chemical properties of the asset. characteristics, the expected life of the asset, the surface of the physical asset, the price of the physical asset, the condition of the physical asset, the location of the physical asset, and/or other attributes, as well as the digital twin and/or can be used to populate the digital twin. Other relevant data sources (e.g., data sources within the management platform described in this application, or external data sources that may affect the characteristics expressed in the digital twin (e.g., how ambient air pressure or temperature expands or contracts) may include identifiers of other digital twins contained within or associated with environmental data sources that may affect the physical dimensions of the asset. Examples of the behavior of a physical asset include the physical asset's state of matter (e.g., solid, liquid, plasma, or gas), the melting point of the physical asset, the density of the physical asset when in a liquid state, and the physical asset's physical state when in a liquid state. Viscosity of an asset, freezing point of a physical asset, density of a physical asset when in a solid state, hardness of a physical asset when in a solid state, malleability of a physical asset, buoyancy of a physical asset, conductivity of a physical asset, Electromagnetic properties, radiation properties, optical properties (e.g., reflectivity, transparency, opacity, albedo, etc.), wave interaction properties (e.g., transparency or opacity to radio waves, reflection properties, shielding properties, etc.), and physical property properties. This can include things like combustion point, how humidity affects a physical asset, how water or other liquids affect a physical asset, and more. In another example, the set of characteristics of the device may include the type of device, the dimensions of the device, the mass of the device, the density of the device, the material(s) of the device, the physical properties of the material(s), the surface of the device, and the output of the device. , the state of the device, the location of the device, the trajectory of the device, the identifier of another digital twin to which the device is connected and/or contains, etc. Examples of the behavior of the device include: maximum acceleration of the device, maximum speed of the device, possible motions of the device, possible configurations of the device, operating modes of the device, heating profile of the device, cooling profile of the device, processes performed by the device, It may include operations performed by, etc. Exemplary characteristics of the environment may include environmental dimensions, environmental air pressure, environmental temperature, environmental humidity, environmental air currents, physical objects within the environment, environmental currents (in the case of a body of water), etc. Examples of the behavior of the environment may include scientific laws that govern the environment, processes performed in the environment, rules or regulations that must be followed in the environment, etc.

In embodiments, the characteristics of the digital twin may be adjusted. For example, the temperature of the digital twin, the humidity of the digital twin, the geometry of the digital twin, the material of the digital twin, the dimensions of the digital twin, or any other suitable parameter may be used to determine the current state data and/or the predicted state of the corresponding entity. It can be adjusted to suit.

In embodiments, a digital twin may be rendered by a computing device so that a human user can view a digital representation of a set of physical assets, devices, or other entities, and/or their environment. For example, the digital twin can be rendered and provided as output, or the output can be provided to a display device. In some embodiments, the digital twin may be rendered and output in augmented reality and/or virtual reality displays. For example, a user can view a 3D rendering of the environment (e.g., using a monitor or virtual reality headset). While doing so, users can inspect digital twins of physical assets or devices within the environment. In embodiments, a user may view processes (e.g., inventorying, loading, packing, shipping, etc.) performed on one or more digital twins. In embodiments, a user may provide input to control one or more characteristics of the digital twin through a graphical user interface.

In some embodiments, adaptive intelligence system layer 614 is configured to run simulations using a digital twin. For example, adaptive intelligence system layer 614 may iteratively adjust one or more parameters of the digital twin and/or one or more embedded digital twins. In an embodiment, adaptive intelligence system layer 614 may, for each set of parameters, run a simulation based on the set of parameters and collect simulation result data resulting from the simulation. In other words, the adaptive intelligence system layer 614 may collect any results resulting from the simulation as well as the digital twin used during the simulation and the characteristics of the digital twin, within or including the digital twin. For example, in running a simulation on a digital twin of a shipping container, adaptive intelligence system layer 614 can vary the materials of the shipping container and run the simulation resulting from different combinations. In this example, the outcome may be whether the goods contained in the shipping container reach their destination undamaged. During the simulation, the adaptive intelligence system layer 614 determines the external temperature of the container (e.g., the temperature characteristics of a digital twin of the container's environment may be adjusted between or during simulations), the dimensions of the container, and the products inside the container. (represented by the digital twin of the product), the motion of the container, the humidity inside the container, and/or any other characteristics of the container, the environment, and/or the contents within the container. For each simulation instance, adaptive intelligence system layer 614 may record the parameters used to perform the simulation instance and the results of the simulation instance. In embodiments, each digital twin may include, reference, or be associated with a set of physical constraints that define boundary conditions for simulation. For example, the physical limitations of a digital twin in an outdoor environment include the gravitational constant (e.g., 9.8 m/s2), maximum temperature (e.g., 60 degrees Celsius), and minimum temperature (e.g., -80 degrees Celsius). , maximum humidity (e.g., 110% humidity), coefficient of friction of the surface, maximum speed of the object, maximum salinity of water, maximum acidity of water, and minimum acidity of water. Additionally or alternatively, simulations may adhere to scientific formulas, such as reflecting principles or laws of physics, chemistry, materials science, biology, geometry, etc. For example, simulations of the physical behavior of an object may comply with the laws of thermodynamics, laws of motion, laws of fluid dynamics, laws of buoyancy, laws of heat transfer, laws of cooling, etc. Accordingly, when the adaptive intelligence system layer 614 performs a simulation, the simulation can comply with physical constraints and scientific laws, so that the results of the simulation mimic real-world results. Results from the simulation will be compared with real-world data (e.g., measured characteristics of the container, environment of the container, contents of the container, and resulting results) and presented to human users to ensure convergence of the digital twin with the real world. may be used, and/or may be used to train a machine learning model.

38 illustrates an example embodiment of a system for controlling and/or performing decisions, predictions, and/or classifications on behalf of value chain system 2030. In embodiments, artificial intelligence system 2010 utilizes one or more machine learning models 2004 to perform value chain-related tasks on behalf of value chain system 2030 and/or make decisions on behalf of value chain system 2030. , classification, and/or prediction. In some embodiments, machine learning system 2002 trains machine learned model 2004 based on training data 2062, results data 2060, and/or simulation data 2022. As used in this application, the term machine learning model may refer to any suitable type of model that is learned in a supervised, unsupervised, or hybrid manner. Examples of machine learning models include neural networks (e.g., deep neural networks, convolutional neural networks, etc.), regression-based models, decision trees, hidden forests, hidden Markov models, Bayesian models, etc. In embodiments, artificial intelligence system 2010 and/or value chain system 2030 may interact with one or more machine learning models and decisions made by artificial intelligence system 2010 based in part on inputs to the models (e.g. , decision, classification, and prediction) can be provided to the machine learning system (2002). The machine learning system can in turn strengthen/retrain the machine learned model (2004) based on the feedback. Additionally, in embodiments, machine learning system 2002 may train a machine learning model based on simulation data 2022 generated by digital twin simulation system 2020. In such embodiments, digital twin simulation system 2020 may be instructed to run specific simulations using one or more digital twins representing objects and/or environments that are managed, maintained, and/or monitored by the value chain system. . In this way, the digital twin simulation system (2020) can provide a richer data set that the machine learning system (2002) can use to train/strengthen the machine learned model. Additionally or alternatively, the digital twin simulation system 2020 may be utilized by the artificial intelligence system 2010 to test decisions made by the artificial intelligence system 2010 before providing the decisions to the value chain entity.

In the illustrated example, machine learning system 2002 may receive training data 2062, results data 2060, and/or simulation data 2022. In embodiments, training data may be data used to initially train a model. Training data may be provided by domain experts, collected from various data sources, and/or obtained from historical records and/or scientific experiments. Training data 2062 may include quantified characteristics of an item or environment and results related to the quantified characteristics. In some embodiments, training data may be structured as n-tuples, where each tuple includes an outcome and a respective set of features related to the outcome. In embodiments, resulting data 2060 includes real-world data (e.g., data measured or captured from one or more of IoT sensors, value chain entities, and/or other sources). Outcome data may include outcomes and characteristics associated with the outcomes. The resulting data may be provided by the value chain system 2030 utilizing the artificial intelligence system 2010 and/or other data sources during the operation of the value chain entity system 2010. Whenever an outcome (whether negative or positive) is realized, the Value Chain Entity System (2010), Artificial Intelligence System (2010), as well as any other data source (2050), collects data about the outcome into a Machine Learning System (2002). It can be printed to . In embodiments, this data may be provided to a machine learning system via an API in the adaptive intelligence system layer 614. Additionally, in embodiments, the adaptive intelligence system layer 614 may obtain data from other types of external data sources that may provide insightful data, but not necessarily value chain entities. For example, weather data, stock market data, news events, etc. may be collected, crawled, subscribed to, etc. to supplement the results data (and/or training data and/or simulation data).

In some embodiments, machine learning system 2002 may receive simulation data 2022 from digital twin simulation system 2020. Simulation data 2022 may be arbitrary data relating to a simulation using a digital twin. Simulation data 2022 may be similar to result data 2060, but the results are simulated results from a run simulation rather than real-world data. In embodiments, simulation data 2022 may include characteristics of and results resulting from the digital twin and any other digital twin used to perform the simulation. In embodiments, the digital twin simulation system 2020 may iteratively adjust characteristics of a digital twin as well as other digital twins that include or include the digital twin. During each iteration, digital twin simulation system 2020 may provide characteristics of the simulation (e.g., characteristics of all digital twins involved in the simulation) to artificial intelligence system 2010, and the artificial intelligence system may then provide Predictions, classifications, or any other decisions are output to the Digital Twin Simulation System (2020). The digital twin simulation system 2020 may use decisions from the artificial intelligence system 2010 to run a simulation (which may result in a series of decisions resulting from state changes in the simulation). In each iteration, the Digital Twin Simulation System (2020) determines the characteristics used to run the simulation, arbitrary decisions from the artificial intelligence system (2010) used by the Digital Twin Simulation System (2020), and machine learning from the simulation. The results to system 2002 may be output to machine learning system 2002 such that the characteristics, decisions, and results of the simulation are used to further train the model(s) used by the artificial intelligence system during the simulation. .

In some embodiments, training data, results data 2060, and/or simulation data 2022 may be fed into a data lake (e.g., a Hadoop data lake). A machine learning system (2002) can structure data from a data lake. In embodiments, machine learning system 2002 may use the collected data to train/strengthen a model to improve the accuracy of the model (e.g., minimize the model's error value). A machine learning system may execute a machine learning algorithm on collected data (e.g., training data, result data, and/or simulation data) to obtain a model. Depending on the type of model, the machine learning algorithm will be different. Examples of learning algorithms/models include (e.g., deep neural networks, convolutional neural networks, etc., as described throughout this disclosure), statistical models (e.g., regression-based models, etc.), decision trees and other decision models; Includes random/hidden forest, hidden Markov model, Bayesian model, etc. When collecting data from the digital twin simulation system 2020, the machine learning system 2002 can train models for scenarios that the value chain system 2030 has not yet faced. In this way, the resulting model will have less “unexplored” feature space, which can lead to improved decisions by artificial intelligence systems (2010). Additionally, because the digital twin is based in part on assumptions, the characteristics of the digital twin can be updated/corrected when real-world behavior differs from that of the digital twin. An example is provided below.

39 illustrates an example of a container fleet management system 2070 interfacing with an adaptive intelligence system layer 614. In an example embodiment, container fleet management system 2070 may be configured to automate one or more aspects of the value chain as it applies to containers and shipping. In embodiments, container fleet management system 2070 may include one or more software modules executed by one or more server devices. These software modules can select the containers to use (e.g., size of container, type of container, provider of containers, etc.) for a set of one or more shipments, schedule delivery/pickup of the containers, select a shipping route, and Can be configured to determine the type of storage for the containers (e.g., outdoor or indoor), select a location for each container while awaiting delivery, and manage bills of lading and/or other suitable container fleet management tasks. there is. In embodiments, machine learning system 2002 trains one or more models that are utilized by artificial intelligence system 2010 to make classifications, predictions, and/or other decisions related to container fleet management. In an exemplary embodiment, the model (2004) determines that a container, given one or more task-related characteristics, can Trained to select the type of As such, the machine learning system 2002 can train a model using n-tuples containing task-related features for a particular event and one or more results associated with the particular event. In this example, the task-related characteristics for a particular event (e.g., a shipment) include the type of container used, the contents of the container, the characteristics of the container contents (e.g., cost, perishability, temperature limits, etc.), and the type of container used. This may include, but is not limited to, source and destination, whether the containers are being shipped via truck, rail, or ship, time of year, cost of each container, and/or other relevant characteristics. In this example, the consequences associated with a particular event may be whether the contents arrived safely, replacement costs associated with any damage or loss (if any), total shipping time, and/or total shipping costs (e.g., whether shipping a container (how much it costs) Additionally, since international and/or interstate logistics may involve many different sources, destinations, contents, weather conditions, etc., simulations simulating different shipping events can be run to enrich the data used to train the model. there is. For example, simulations can be run for different combinations of ports and/or train depots for different combinations of sources, destinations, products, and times of year. In this example, different digital twins may be created to represent different combinations (e.g., digital twins of products, containers, and shipping-related environments), whereby one or more characteristics of the digital twin can be modified for different simulations. The results of each simulation can be changed and recorded in a tuple with appropriateness. In this way, models can be trained for specific combinations of routes, content, times of year, container types, and/or costs that may not have been previously encountered in real-world outcome data. Another example of training a container fleet management model is to know the contents of the container, when the container needs to be moved, and where the container should be stored in a storage facility given the type of container, location, time of year, etc. (e.g. stack may include a model that is trained to determine (interior, indoor, or outdoor, etc.)

In operation, artificial intelligence system 2010 may use the model 2004 discussed above to make container fleet management decisions on behalf of container fleet management system 2070, given one or more characteristics associated with a task or event. For example, the artificial intelligence system 2010 may select the type of container to use for a particular shipment (e.g., the material of the container, the dimensions of the container, the brand of the container, etc.). In this example, container fleet management system 2070 may provide characteristics of upcoming shipments to artificial intelligence system 2010. These characteristics include what is being shipped (e.g., the type(s) of product within the shipment), the size, source and destination of the shipment, the date by which the shipment must be delivered, and/or the desired date or range of dates for delivery. can do. In embodiments, artificial intelligence system 2010 may feed these features to one or more of the models discussed above to obtain one or more decisions. These decisions may include which type of container to use and/or which shipping route to use, whereby the decision is to minimize overall shipping costs (e.g., costs for container and transportation plus any alternative costs). may be chosen to do so. Container fleet management system 2070 may then use the decision(s) made by artificial intelligence system 2010 to initiate a delivery event. Additionally, after a delivery event, the results of the event (e.g., total delivery time, any reported damage or loss, replacement cost, total cost) are analyzed by machine learning systems (2002) to strengthen the models used to make decisions. can be reported to Additionally, in some embodiments, the output of container fleet management system 2070 and/or other value chain entity data source 2050 may be used to update one or more characteristics of one or more digital twins via digital twin system 2020. You can.

40 illustrates an example of a logistics design system interfacing with the adaptive intelligence system layer 614. In embodiments, a logistics design system may be configured to design one or more aspects of a logistics solution. For example, a logistics design system may be configured to receive one or more logistics factors (e.g., from a user via a GUI). In embodiments, the logistics factors may include one or more current conditions, past conditions, or future conditions of the organization (or potential organization) involved in forming the logistics solution. Examples of logistics factors include the type(s) of product being produced/grown/shipped, characteristics of those products (e.g. dimensions, weight, shipping requirements, shelf life, etc.), location of manufacturing sites, location of distribution facilities, warehouse May include house location, location of customer base, market penetration in specific areas, expansion location, supply chain characteristics (e.g. parts/supplies/resources required, suppliers, supplier location, buyer, purchaser location), etc. However, without being limited thereto, one or more design recommendations may be determined based on these factors. Examples of design recommendations include supply chain recommendations (e.g., proposed suppliers (e.g., resource or parts suppliers), implementation of a smart inventory system to order parts on-demand from available suppliers, etc.), storage and transportation recommendations, etc. (e.g., proposed delivery routes, proposed delivery types (e.g., air, freight, truck, ship), proposed storage developments (e.g., location and/or dimensions of new warehouse), infrastructure Recommendations may include (e.g., updates to machines, addition of cold storage, addition of hot storage, etc.), and combinations thereof.In embodiments, the logistics design system determines recommendations to optimize the results. Results Examples of optimization may include manufacturing time, manufacturing cost, delivery time, delivery cost, loss rate, environmental impact, compliance with a set of rules/regulations, etc. Examples of optimization include increased production throughput, reduced production cost, reduction These include reduced shipping costs, reduced delivery times, reduced carbon footprint, and combinations of these.

In embodiments, the logistics design system may interface with the artificial intelligence system 2010 to provide logistics factors and receive design recommendations based thereon. In embodiments, artificial intelligence system 2010 may utilize one or more machine learning models 2004 (e.g., a logistics design recommendation model) to determine recommendations. As will be discussed, a logistic design recommendation model can be trained to optimize one or more outcomes given a set of logistic factors. For example, a logistics design recommendation model trained to design a supply chain may identify a set of suppliers that can supply a given manufacturer, the manufacturer's location, needed supplies, and/or other factors. The set of suppliers can then be used to implement on-demand supply-side inventory. In another example, a logistics design recommendation may take the same features from different manufacturers and recommend the purchase and use of one or more 3D printers.

In embodiments, the artificial intelligence system 2010 utilizes the digital twin system 2020 to create a digital twin of a logistics system implementing logistics design recommendations (and, in some embodiments, alternative systems implementing other design recommendations). can be created. In this embodiment, digital twin system 1700 can receive design recommendations and create a digital twin of the logistics environment that mirrors the recommendations. In embodiments, artificial intelligence system 2010 may utilize a digital twin of a logistics environment to run simulations for proposed solutions. In an embodiment, the digital twin system 1700 may display a digital twin of the logistics environment to a user through a display device (e.g., a monitor or VR headset). In embodiments, users can view simulations in a digital twin. Additionally, in embodiments, digital twin system 1700 may provide a graphical user interface with which a user can interact to adjust the design of the logistics environment to adjust the design. The design provided (at least in part) by the user can also be represented in the digital twin of the logistics environment, whereby the digital twin system 2020 can perform simulations using the digital twin.

In some embodiments, simulations run by digital twin system 1700 may be used to train a recommendation model. Additionally, when a design recommendation is implemented by an organization, the organization's logistics system is configured to report resulting data corresponding to the design recommendation to the machine learning system 2002 (e.g., through sensors, computing devices, manual human input). can be configured, and machine learning systems can use the resulting data to strengthen logistics design recommendation models.

41 illustrates an example of a packaging design system interfacing with the adaptive intelligence system layer 614. In embodiments, a packaging design system may be configured to design one or more aspects of packaging for physical objects transported in a value chain network. In some embodiments, the packaging design system may select one or more packaging attributes (e.g., size, material, padding, etc.) of the packaging to optimize one or more outcomes associated with transportation of the physical object. For example, packaging properties may be selected to reduce cost, reduce loss/damage, reduce weight, reduce plastic or other non-biodegradable waste, etc. In an embodiment, the packaging design system utilizes an artificial intelligence system (2010) to obtain packaging attribute recommendations. In embodiments, a packaging design system may provide one or more characteristics of a physical object. In embodiments, characteristics of the physical object may include dimensions of the physical object, mass of the physical object, source of the physical object, one or more potential destinations of the physical object, manner in which the physical object is shipped, etc. In embodiments, the packaging design system may further provide one or more optimization goals for package design (e.g., reduce cost, reduce damage, reduce environmental impact). In response, artificial intelligence system 2010 may determine one or more recommended packaging attributes based on physical asset characteristics and given goals. In an embodiment, a packaging design system receives packaging attributes and creates a package design based thereon. Package design may include materials to be used, external dimensions of the packaging, internal dimensions of the packaging, shape of the packaging, padding/stuffing for the packaging, etc.

In some embodiments, the packaging design system may provide the packaging design to a digital twin system 2020 that creates a digital twin of the packaging and physical assets based on the packaging design. Digital twins of packaging and physical assets can be used to run simulations that test the packaging (e.g., will the packaging hold up during shipping, will the packaging provide adequate insulation/padding, etc.) In embodiments, the results of the simulation may be returned to the packaging design system, and the packaging design system may output the results to a user. In embodiments, a user can accept the packaging design, adjust the packaging design, or reject the design. In some embodiments, the digital twin system may run simulations on one or more digital twins to test different conditions the package may experience (e.g., outside in the snow, being rocked on a boat, being moved by a forklift, etc.). You can. In some embodiments, the digital twin system may output the results of the simulation to a machine learning system 2002, which trains a packaging design model based on the characteristics used to run the simulation and the results resulting therefrom. /Can be strengthened.

In embodiments, machine learning system 2002 may receive resulting data from packaging design systems and/or other value chain entity data sources (e.g., smart warehouse, user feedback, etc.). A machine learning system (2002) can use this resulting data to train/strengthen a packaging design model. Additionally, in some embodiments, the resulting data may be used by the Digital Twin System (2020) to update/correct any inaccurate assumptions used by the Digital Twin System (e.g., flexibility of the packaging material, water resistance of the packaging material, etc.). It can be used by .

42 illustrates an example of a waste mitigation system interfacing with the adaptive intelligence system layer 614. In embodiments, a waste mitigation system may utilize processes within the value chain (e.g., For example, it is designed to analyze product manufacturing, oil refining, fertilization, water treatment, etc.). In embodiments, the waste mitigation system may interface with the artificial intelligence system 2010 to automate one or more processes for mitigating waste.

In embodiments, artificial intelligence system 2010 may provide control decisions to a waste mitigation system to mitigate solid waste generation. Examples of waste generation may include excess plastic or other non-biodegradable waste, hazardous or toxic waste (e.g., nuclear waste, petroleum coke, etc.), etc. In some of these embodiments, artificial intelligence system 2010 may receive one or more characteristics of a process (or “process characteristics”). Examples of process characteristics may include, but are not limited to, steps in the process, materials used, properties of materials used, etc. An artificial intelligence system (2010) may utilize one or more machine learning models to control a process. In embodiments, a machine learned model may be trained to classify waste conditions and/or causes of waste conditions. In some of these embodiments, the artificial intelligence system 2010 may determine or select a waste mitigation solution based on classified waste conditions. For example, in some embodiments, artificial intelligence system 2010 may apply rule-based logic to determine adjustments to make to a process to reduce or address waste conditions. Additionally or alternatively, artificial intelligence can leverage models to recommend adjustments to make to processes to reduce or address waste conditions.

In embodiments, artificial intelligence system 2010 may utilize digital twin system 2020 to mitigate waste generated by a process. In embodiments, digital twin system 2020 may run an iterative simulation of a process on a digital twin of the environment in which the process is performed. As the simulation runs, the artificial intelligence system 2010 may monitor the results of the simulation to determine waste conditions and/or causes of waste conditions. During the simulation, the artificial intelligence system 2010 may adjust one or more aspects of the process to determine whether the adjustments alleviated waste conditions, worsened waste conditions, or were ineffective. When adjustments are found to alleviate waste conditions, the artificial intelligence system 2010 can adjust other aspects of the process to determine whether improvements can be realized. In embodiments, artificial intelligence system 2010 may perform genetic algorithms when iteratively adjusting aspects of a process in a digital twin simulation. In this embodiment, artificial intelligence system 2010 may identify aspects of the process that can be adjusted to mitigate waste generation.

Smart Project Management Facilities

43, an embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). a set, a set of adaptive intelligence facilities or adaptive intelligence systems 614 (including artificial intelligence 1160), a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. You can use micro-service architecture. Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

In an embodiment, the adaptive intelligence system layer 614 may configure a set of value chain project management tasks, a set of demand management applications 1502, a set of supply chain applications 1500, and intelligent products based on processing current state information. Automated recommendations for a set of application outputs and/or results 1040 for a set of applications 1510, a set of asset management applications 1530, and a set of enterprise resource management applications 1520 for categories of products. It may further include a set of automated project management facilities 21006 that provide.

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; of value chain project management tasks based on processing a set of current status information and results for a set of demand management applications, a set of supply chain applications, a set of intelligent product applications and a set of enterprise resource planning applications for a category of goods. It may include a set of project management facilities that provide automated recommendations for the set.

In an embodiment, the set of project management facilities may include procurement projects, logistics projects, reverse logistics projects, fulfillment projects, distribution projects, warehousing projects, inventory management projects, product design projects, product management projects, shipping projects, maritime projects, and loading. Or, it is configured to manage many different types of projects, such as unloading projects, packing projects, purchasing projects, marketing projects, sales projects, analysis projects, demand management projects, demand planning projects, resource planning projects, etc.

In embodiments, a project management facility is configured to manage a set of procurement projects. In embodiments, a project management facility is configured to manage a set of logistics projects. In embodiments, a project management facility is configured to manage a set of reverse logistics projects. In embodiments, a project management facility is configured to manage a set of fulfillment projects.

In embodiments, a project management facility is configured to manage a set of distribution projects. In embodiments, a project management facility is configured to manage a set of warehousing projects. In embodiments, a project management facility is configured to manage a set of inventory management projects. In embodiments, a project management facility is configured to manage a set of product design projects.

In embodiments, a project management facility is configured to manage a set of product management projects. In embodiments, a project management facility is configured to manage a set of delivery projects. In embodiments, a project management facility is configured to manage a set of offshore projects. In embodiments, a project management facility is configured to manage a set of loading or unloading projects.

In embodiments, a project management facility is configured to manage a set of packing projects. In embodiments, a project management facility is configured to manage a set of purchasing projects. In embodiments, a project management facility is configured to manage a set of marketing projects. In embodiments, a project management facility is configured to manage a set of sales projects.

In embodiments, a project management facility is configured to manage a set of analysis projects. In embodiments, a project management facility is configured to manage a set of demand management projects. In embodiments, a project management facility is configured to manage a set of demand planning projects. In embodiments, a project management facility is configured to manage a set of resource planning projects.

Smart Task Recommendations

45, an embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). a set, a set of adaptive intelligence facilities or adaptive intelligence systems 614 (including artificial intelligence 1160), a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. You can use micro-service architecture.

Platform 604 includes applications 614 (processes, workflows, , activities, events, use cases, and applications).

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; of value chain project management tasks based on processing a set of current status information and results for a set of demand management applications, a set of supply chain applications, a set of intelligent product applications and a set of enterprise resource planning applications for a category of goods. It may include a set of project management facilities that provide automated recommendations for the set.

In an embodiment, adaptive intelligence system layer 614 provides automated recommendations for a set of value chain process tasks 1710 that provides automated recommendations for a set of value chain processes based on processing current state information. a set of process automation facilities 1710, a set of application outputs and/or results 1040 for a set of demand management applications 1502, a set of supply chain applications 1500, and an intelligent product application 1510. It may further include a set of, a set of asset management applications 1530 and a set of enterprise resource management applications 1520 for categories of products. In some examples, process automation facility 1710 may be used with basic rule-based training and recommendations. This may involve following a set of rules outlined by an expert, such as performing a task when a trigger occurs. In another example, process automation facility 1710 may use deep learning to observe interactions, such as deep learning on outcomes, to learn to recommend decisions or tasks that produce the highest return on investment (ROI) or other outcome-based yield. Available. Process automation facility 1710 may be used to provide collaborative filtering, for example, to find the set of experts who are most similar in terms of which work performed and tasks completed are most similar. For example, the base software could be used to find customers who are similar to another set of customers to sell to, make different offers, or change prices accordingly. In general, given a set of basic pattern data for a customer segment in context, purchasing patterns, such as knowledge of cost and pricing patterns for that customer, can be determined for that customer segment. This information could ideally be based on deep learning or rules or collaborative filtering type operations that attempt to leverage similar decisions made by similarly situated people (e.g. recommending a movie to a cohort of similar people). It can be used to learn to focus the next set of activities on pricing, promotion, and demand management toward value.

In embodiments, a set of facilities that provide automated recommendations for a set of value chain process tasks may include, without limitation, a product configuration activity, a product selection activity for a customer, a supplier selection activity, a shipper selection activity, a route selection activity, It provides recommendations involving a wide range of types of activities, such as factory selection activities, product type classification activities, product management activities, logistics activities, reverse logistics activities, artificial intelligence configuration activities, maintenance activities, product support activities, product recommendation activities, etc.

In embodiments, automated recommendations relate to a set of product configuration activities. In embodiments, automated recommendations involve a set of product selection activities for a customer. In embodiments, automated recommendations relate to a set of supplier selection activities. In embodiments, automated recommendations relate to a set of courier selection activities.

In embodiments, automated recommendations relate to a set of route selection activities. In embodiments, automated recommendations involve a set of factory selection activities. In embodiments, automated recommendations relate to a set of sorting activities by product type. In embodiments, automated recommendations relate to a set of product management activities. In an embodiment, the automated recommendation relates to a set of logistics activities.

In an embodiment, the automated recommendation relates to a set of reverse logistics activities. In embodiments, automated recommendations relate to a set of artificial intelligence-constructed activities. In embodiments, automated recommendations relate to a set of maintenance activities. In embodiments, automated recommendations relate to a set of product support activities. In embodiments, automated recommendations involve a set of product recommendation activities.

Optimized routing among nodes

44, an embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). a set, a set of adaptive intelligence facilities or adaptive intelligence systems 614 (including artificial intelligence 1160), a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. You can use micro-service architecture. Platform 604 is a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; A cloud-based management platform for value chain networks with adaptive intelligence facilities, data storage facilities, and monitoring facilities; and a set of applications to enable an enterprise to manage a set of value chain network entities from point of origin to point of customer use; A set of routing facilities generates a set of routing instructions for routing information between a set of nodes within a value chain network based on current state information about the value chain network.

In embodiments, adaptive intelligence system layer 614 may, for example, process current state information 1730, set of application outputs and/or results 1040, or collected by or used by VCNP 102. It may further include a set of routing facilities 1720 that generate a set of routing instructions for routing information between sets of nodes in the value chain network, based on other information being received. Routing includes a set of demand management applications 1502, a set of supply chain applications 1500, a set of intelligent product applications 1510, a set of asset management applications 1530, and enterprise resource management applications 1520 for categories of goods. May include routing for the benefit of the set.

In an embodiment, a set of routing facilities that generate a set of routing instructions for routing information between a set of nodes within a value chain network include priority-based routing, master controller routing, least-cost routing, rule-based routing, and genetically programmed routing. routing, random linear network coding routing, traffic-based routing, spectrum-based routing, RF condition-based routing, energy-based routing, latency-sensitive routing, protocol compatibility-based routing, dynamic spectrum access routing, peer-to-peer negotiated routing, queue-based routing. A wide variety of routing systems or configurations are used, such as, but not limited to, routing, etc.

In an embodiment, routing includes priority based routing. In an embodiment, routing includes master controller routing. In an embodiment, routing includes least cost routing. In an embodiment, routing includes rule-based routing. In an embodiment, routing includes genetically programmed routing.

In an embodiment, the routing includes random linear network coding routing. In an embodiment, routing includes traffic based routing. In an embodiment, routing includes spectrum-based routing.

In an embodiment, routing includes RF condition based routing. In an embodiment, routing includes energy-based routing. In an embodiment, routing includes latency-sensitive routing.

In an embodiment, routing includes protocol compatibility based routing.

In an embodiment, routing includes dynamic spectrum access routing. In an embodiment, routing includes peer-to-peer negotiated routing. In an embodiment, routing includes queue-based routing.

In embodiments, state information for a value chain network may include, but is not limited to, traffic state, congestion state, bandwidth state, operational state, workflow progress state, incident state, damage state, safety state, power availability state, worker state, data state, Availability Status, Predicted System Status, Delivery Location Status, Delivery Timing Status, Delivery Status, Expected Delivery Status, Environmental Conditions Status, System Diagnostic Status, System Fault Status, Cybersecurity Status, Compliance Status, Demand Status, Supply Status, Price It involves a wide range of states, events, workflows, activities, occurrences, etc., such as states, volatile states, need states, states of interest, aggregated states for groups or populations, individual states, etc.

In an embodiment, the state information involves traffic status. In an embodiment, the state information involves congestion conditions. In an embodiment, the status information involves bandwidth status. In embodiments, the status information involves operational status. In embodiments, the status information involves workflow progress.

In embodiments, the status information involves event status. In an embodiment, the status information involves a corruption condition. In an embodiment, the status information entails a safety status.

In an embodiment, the status information involves power availability status. In an embodiment, the status information involves operator status. In an embodiment, the status information involves data availability status.

In embodiments, the state information involves predicted system states. In embodiments, the status information involves delivery location status. In embodiments, the status information involves delivery timing status. In an embodiment, the status information entails delivery status.

In embodiments, the status information entails expected delivery status. In embodiments, the status information involves environmental condition status.

In embodiments, the status information involves system diagnostic status. In an embodiment, the status information involves a system fault condition. In embodiments, the status information involves cybersecurity status. In embodiments, the status information involves compliance status.

DASHBOARD FOR MANAGING DIGITAL TWINS

47, an embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). a set, a set of adaptive intelligence facilities or adaptive intelligence systems 614 (including artificial intelligence 1160), a set of data storage facilities or systems 624, and a set of monitoring facilities or systems 808. You can use micro-service architecture. Platform 604 includes a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; and a dashboard for managing the set of digital twins, wherein at least one digital twin represents a set of supply chain entities, workflows, and assets, and at least one other digital twin represents a set of demand management entities and workflows. represents.

In an embodiment, VCNP 604 may further include a dashboard 4200 for managing the set of digital twins 1700. In embodiments, this includes different twins, such as where one digital twin 1700 represents a set of supply chain entities, workflows, and assets and the other digital twin 1700 represents a set of demand management entities and workflows. can do. In some example embodiments, managing a set of digital twins 1700 may refer to a configuration as described in this disclosure (e.g., via dashboard 4200). For example, digital twin 1700 establishes and manages the enterprise digital twin and the enterprise's associated metadata, configures the data structures and data listening threads that operate the enterprise digital twin, and includes access characteristics, processing characteristics, and automation characteristics. , can be configured using a digital twin configuration system to configure the characteristics of the enterprise digital twin, including reporting features, etc., each of which can be configured (e.g., the role(s) the enterprise digital twin serves, the entities it represents, This can be influenced by the type of enterprise digital twin (based on the workflows it supports or enables, etc.). In an example embodiment, a digital twin configuration system may receive the types of digital twins that can be supported for an enterprise, as well as the different objects, entities, and/or states to be represented in each type of digital twin. . For each type of digital twin, the digital twin configuration system can determine one or more data sources and types of data that supply or otherwise support each object, entity, or state that appears in each type of digital twin. and any internal or external software request (e.g., API call) that obtains the identified data type, or other suitable, such as a webhook, that can be configured to automatically receive data from an internal or external data source. The data acquisition mechanism can be determined. In some embodiments, the digital twin configuration system may determine internal and/or external software requests that support the identified data types by analyzing the relationships between different types of data and their granularity corresponding to specific states/entities/objects. . Additionally or alternatively, the user may define (e.g., via a GUI) data sources and/or software requests and/or other data acquisition mechanisms that support each data type displayed in each digital twin. . In this example embodiment, a user can indicate the data sources that can be accessed and the type of data to be obtained from each data source.

The dashboard may include, for example, a set of demand management applications 1502, a set of supply chain applications 1500, a set of intelligent product applications 1510, a set of asset management applications 1530, and enterprise resource management for categories of goods. may be used to construct a digital twin 1700 for use in the collection, processing, and/or presentation of information collected on platform 604, such as state information 1730, for the benefit of a set of applications 1520. You can.

In an embodiment, a dashboard is for managing a set of digital twins, wherein at least one digital twin represents a set of supply chain entities and workflows, and at least one other digital twin represents a set of demand management entities and workflows. Represents a set.

In an embodiment, the entities and workflows are related to a set of enterprise products. In an embodiment, the entities and workflows are associated with a set of enterprise suppliers. In an embodiment, the entities and workflows are associated with a set of producers of a set of products. In an embodiment, the entities and workflows relate to a set of manufacturers of a set of products.

In an embodiment, the entities and workflows relate to a retailer's set of lines of products. In an embodiment, the entities and workflows relate to a set of operators involved in the ecosystem for a category of products. In an embodiment, the entities and workflows relate to a set of owners of a set of assets involved in the value chain for a set of products. In an embodiment, the entities and workflows relate to an operator's set of assets involved in a value chain for a set of products.

In embodiments, entities and workflows are associated with a set of operational facilities. In embodiments, entities and workflows are associated with sets of customers. In embodiments, entities and workflows are associated with sets of consumers. In embodiments, entities and workflows are associated with sets of workers.

In embodiments, entities and workflows relate to a set of mobile devices. In an embodiment, entities and workflows relate to a set of wearable devices. In an embodiment, the entities and workflows are associated with a set of distributors. In an embodiment, entities and workflows are associated with a set of resellers.

In embodiments, entities and workflows are associated with a set of supply chain infrastructure facilities. In embodiments, entities and workflows are associated with a set of supply chain processes. In embodiments, entities and workflows are associated with a set of logistics processes. In embodiments, entities and workflows are associated with a set of reverse logistics processes.

In embodiments, entities and workflows are related to a set of demand forecasting processes. In embodiments, entities and workflows are associated with a set of demand management processes. In embodiments, entities and workflows are associated with a set of demand aggregation processes. In an embodiment, entities and workflows relate to sets of machines.

In an embodiment, entities and workflows are associated with a set of vessels. In an embodiment, the entities and workflows are associated with a set of barges. In an embodiment, entities and workflows are associated with a set of warehouses. In an embodiment, the entities and workflows are associated with a set of maritime ports.

In an embodiment, the entities and workflows are associated with a set of airports. In an embodiment, entities and workflows are related to a set of routes. In an embodiment, entities and workflows are associated with sets of channels. In an embodiment, entities and workflows are related to a set of roads.

In an embodiment, the entities and workflows are related to a set of railroads. In an embodiment, entities and workflows are related to a set of bridges. In an embodiment, entities and workflows are associated with a set of tunnels. In an embodiment, the entities and workflows are associated with a set of online retailers.

In an embodiment, the entities and workflows are associated with a set of e-commerce sites. In an embodiment, entities and workflows are related to a set of demand factors. In an embodiment, entities and workflows are associated with sets of provisioning factors. In embodiments, entities and workflows are associated with a set of delivery systems.

In embodiments, entities and workflows are related to sets of floating assets. In embodiments, entities and workflows are associated with a set of origin points. In an embodiment, entities and workflows are associated with a set of destination points. In an embodiment, entities and workflows are associated with a set of points in the repository.

In embodiments, entities and workflows relate to a set of points of product use. In an embodiment, entities and workflows are associated with a set of networks. In embodiments, entities and workflows relate to a set of information technology systems. In embodiments, entities and workflows relate to a set of software platforms.

In an embodiment, entities and workflows are associated with a set of distribution centers. In embodiments, entities and workflows are associated with a set of fulfillment centers. In an embodiment, entities and workflows are related to sets of containers. In an embodiment, the entities and workflows relate to a set of container handling facilities.

In embodiments, entities and workflows are associated with a set of customs. In embodiments, entities and workflows are associated with a set of export controls. In an embodiment, entities and workflows are associated with a set of boundary controls. In an embodiment, entities and workflows are associated with a set of drones.

In embodiments, entities and workflows are associated with sets of robots. In an embodiment, the entities and workflows are associated with a set of autonomous vehicles. In an embodiment, the entities and workflows relate to a set of transport facilities. In embodiments, entities and workflows relate to sets of drones, robots, and autonomous vehicles. In an embodiment, entities and workflows are associated with sets of channels. In an embodiment, the entities and workflows relate to a set of port infrastructure facilities.

In an embodiment, a set of digital twins may include, for example, and without limitation, a distribution twin, a warehousing twin, a port infrastructure twin, a shipping facility twin, an operations facility twin, a customer twin, a worker twin, a wearable device twin, a portable device Twin, mobile device twin, process twin, machine twin, asset twin, product twin, point of origin twin, point of destination twin, supply factor twin, offshore facility twin, floating asset twin, shipyard twin, fulfillment twin, delivery system twin, demand Factor twin, retailer twin, e-commerce twin, online twin, waterway twin, road twin, road twin, railway twin, aviation facility twin, aircraft twin, ship twin, vehicle twin, train twin, autonomous vehicle twin, robot system twin, It may include drone twins, logistics factor twins, etc.

MICROSERVICES ARCHITECTURE

48, one embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems (614), a set of data storage facilities or systems (624), and a set of monitoring facilities or systems (808). Platform 604 includes a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; It may include a set of microservice layers, including an application layer supporting at least one supply chain application and at least one demand management application, wherein the applications in the application layer include data processing services, data collection services, and data storage services. Use a set of common services between sets.

In an embodiment, VCNP 604 includes a set of demand management applications 1502, a set of supply chain applications 1500, a set of intelligent product applications 1510, a set of asset management applications 1530, and The set of enterprise resource management applications 1520 may further include a set of microservice layers including an application layer supporting at least two applications.

Microservice architecture provides several advantages to platform 604. For example, one advantage is that the platform leverages the creation of improved microservices created by others so that developers only need to define inputs and outputs so that they can use easily adapted services created by others. It could be an ability. Additionally, the use of a microservice architecture can provide the ability to modularize microservices into collections that can be used to accomplish tasks. For example, the goal of determining what is happening in a warehouse is achieved with a variety of microservices with minimal cost, such as vision-based services, a set of regular prompts that can be requested and received, reading of event logs or feeds, etc. It can be. Each of these microservices can be a separate microservice that can be easily plugged in and used. If a particular microservice does not perform effectively, the microservice can be easily replaced by another service with minimal impact on other components within the platform. Other microservices that can be used include recommendation services, collaborative filtering services, deep learning with semi-supervised learning services, etc. Microservice architecture can provide modularity at each stage in building the overall workflow. In an example embodiment, microservices can be built for multiple applications that can be consumed, including shared data streams and any other enabled by the microservices architecture.

Recommend IoT data collection architectures from different sensors and cameras

49, an embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems (1160), a set of data storage facilities or systems (624), and a set of monitoring facilities or systems (808). Platform 604 includes a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; It may include a set of microservice layers, including an application layer supporting at least one supply chain application and at least one demand management application, wherein the microservice layer is a set of microservice layers that collect information about supply chain entities and demand management entities. Includes a data collection layer that collects information from a set of Internet resources.

Additionally, methods, systems, components and other elements for an information technology system are provided in this application, including a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; A machine learning/artificial intelligence system configured to generate recommendations for placing additional sensors/and/or cameras on and/or proximate to a value chain entity, wherein data from the additional sensors and/or cameras The chain is fed into a digital twin that represents a set of entities.

In an embodiment, VCNP 604 may further include a set of microservices, where the microservice layer collects information from a set of Internet of Things resources 1172 to collect information about supply chain entities and demand management entities 652. It includes a monitoring system and data collection system layer 614 with a data collection and management system 640 that collects. Microservices include a set of demand management applications 1502 for a category of goods, a set of supply chain applications 1500, a set of intelligent product applications 1510, a set of asset management applications 1530, and an enterprise resource management application 1520. ) can support a variety of applications from the set.

In an embodiment, platform 604 includes a sensor recommendation system 1750 configured to generate recommendations for placing additional sensors 1462 and/or cameras on and/or proximate to value chain network entity 652. A machine learning/artificial intelligence system 1160 may be further included. For example, in some embodiments, sensor recommendation system 1750 may use load, array of signals, emergency conditions, frequency response, maintenance, diagnostics, etc. to generate recommendations. Data from additional sensors 1462 and/or cameras may be fed into a digital twin 1700 representing a set of value chain entities 652. In embodiments, a set of Internet of Things resources that collect information about supply chain entities and demand management entities may collect information from any of the types of entities described throughout this disclosure and in documents incorporated by reference herein. Collect.

In an embodiment, the set of Internet of Things resources may include a camera system, a lighting system, a motion detection system, a weighting system, an inspection system, a machine vision system, an environmental sensor system, an on-board sensor system, an on-board diagnostic system, an environmental control system, and sensor-enables. Network switching and routing systems, RF sensing systems, magnetic sensing systems, pressure monitoring systems, vibration monitoring systems, temperature monitoring systems, heat flow monitoring systems, biological measurement systems, chemical measurement systems, ultrasonic monitoring systems, radiography systems, LIDAR-based monitoring The system may be of a wide variety of types, such as, without limitation, an access control system, a penetrating wave detection system, a SONAR-based monitoring system, a radar-based monitoring system, a computed tomography system, a magnetic resonance imaging system, a network monitoring system, and the like.

In an embodiment, the set of Internet of Things resources includes a set of camera systems. In an embodiment, the set of Internet of Things resources includes a set of lighting systems. In an embodiment, the set of Internet of Things resources includes a set of machine vision systems. In an embodiment, the set of Internet of Things resources includes a set of motion detection systems.

In an embodiment, the set of Internet of Things resources includes a set of weighting systems. In an embodiment, the set of Internet of Things resources includes a set of inspection systems. In an embodiment, the set of Internet of Things resources includes a set of environmental sensor systems. In an embodiment, the set of Internet of Things resources includes a set of onboard sensor systems.

In an embodiment, the set of Internet of Things resources includes a set of onboard diagnostic systems. In an embodiment, the set of Internet of Things resources includes a set of environmental control systems. In an embodiment, the set of Internet of Things resources includes a set of sensor-enabled network switching and routing systems. In an embodiment, the set of Internet of Things resources includes a set of RF sensing systems. In an embodiment, the set of Internet of Things resources includes a set of magnetic sensing systems.

In an embodiment, the set of Internet of Things resources includes a set of pressure monitoring systems. In an embodiment, the set of Internet of Things resources includes a set of vibration monitoring systems. In an embodiment, the set of Internet of Things resources includes a set of temperature monitoring systems. In an embodiment, the set of Internet of Things resources includes a set of heat flow monitoring systems. In an embodiment, the set of Internet of Things resources includes a set of biological measurement systems.

In an embodiment, the set of Internet of Things resources includes a set of chemical measurement systems. In an embodiment, the set of Internet of Things resources includes a set of ultrasonic monitoring systems. In an embodiment, the set of Internet of Things resources includes a set of radiography systems. In an embodiment, the set of Internet of Things resources includes a set of LIDAR-based monitoring systems. In an embodiment, a set of Internet of Things resources includes a set of access control systems.

In an embodiment, a set of Internet of Things resources includes a set of penetrating wave detection systems. In an embodiment, the set of Internet of Things resources includes a set of SONAR-based monitoring systems. In an embodiment, a set of Internet of Things resources includes a set of radar-based monitoring systems. In an embodiment, the set of Internet of Things resources includes a set of computed tomography systems. In an embodiment, the set of Internet of Things resources includes a set of magnetic resonance imaging systems. In an embodiment, the set of Internet of Things resources includes a set of network monitoring systems.

SOCIAL DATA COLLECTION ARCHITECTURE

50, one embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems (1160), a set of data storage facilities or systems (624), and a set of monitoring facilities or systems (808). Platform 604 includes a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; It may include a set of microservice layers, including an application layer supporting at least one supply chain application and at least one demand management application, wherein the microservice layer is a social service layer that provides information about the supply chain entity and the demand management entity. Includes a data collection layer that collects information from a set of network sources.

In an embodiment, VCNP 604 has a data collection layer (e.g., social data collection facility 1760) that collects information from a set of social network resources MPVC 1708 providing information about supply chain entities and demand management entities. For example, it may further include a set of microservice layers including a monitoring system and data collection system layer 614). The social network data collection facility 1760 includes a set of demand management applications 1502 for a category of goods, a set of supply chain applications 1500, a set of intelligent product applications 1510, a set of asset management applications 1530, and A variety of applications may be supported among the set of enterprise resource management applications 1520. Social network data collection (using the social network data collection facility 1760) may be performed to formulate queries, identify social data sources of relevance, configure APIs for data collection, and use appropriate applications ( 630), and may be facilitated by a social data collection configuration interface, such as for routing data.

CROWDSOURCING DATA COLLECTION ARCHITECTURE

51, one embodiment of platform 604 is provided. As in other embodiments, platform 604 includes various data handling layers 608, network connectivity facilities 642 (which may include or connect to a set of interfaces 702 of the various layers of platform 604). A micro-service architecture may be used having a set of adaptive intelligence facilities or adaptive intelligence systems (1160), a set of data storage facilities or systems (624), and a set of monitoring facilities or systems (808). Platform 604 includes a set of applications 614 (processes, (including workflows, activities, events, use cases, and applications).

Accordingly, the present application provides methods, systems, components and other elements for an information technology system, comprising a micro-service architecture, a set of interfaces, network connectivity facilities, coordinated for monitoring and management of a set of value chain network entities; a cloud-based management platform with adaptive intelligence facilities, data storage facilities, and monitoring facilities; A set of applications to enable enterprises to manage a set of value chain network entities from point of origin to point of customer use; It may include a set of microservice layers, including an application layer supporting at least one supply chain application and at least one demand management application, wherein the microservice layer provides information about the supply chain entity and the demand management entity. Includes a data collection layer that collects information from a set of sourcing resources.

In an embodiment, VCNP 604 includes a monitoring system and data collection system layer 614 with a crowdsourcing facility 1770 that collects information from a set of crowdsourcing resources providing information about supply chain entities and demand management entities. It may further include a set of microservice layers including. The crowdsourcing facility 1770 includes a set of demand management applications 1502 for categories of goods, a set of supply chain applications 1500, a set of intelligent product applications 1510, a set of asset management applications 1530, and enterprise resources. Among the set of management applications 1520, a variety of applications may be supported. Crowdsourcing is facilitated by a crowdsourcing interface, such as formulating queries, setting rewards for information, organizing workflows, determining eligibility for participation, and other elements of crowdsourcing. It can be.

VALUE CHAIN DIGITAL TWIN PROCESSING (DTPT)

Referring now to FIG. 52 , a set of value chain network digital twins 1700 are shown, which represent a set of value chain network entities 652 . Digital twin 1700 is configured to simulate the characteristics, state, operation, behavior and other aspects of value chain network entity 652. Digital twin 1700 may have a visual user interface, for example, in the form of a 3D model, or may consist of a system specification or ontology that describes an architecture including components and interfaces of value chain network entities 652. there is. Digital twin 1700 is a value chain network entity (e.g., captured via sensors, via user input, and/or determined by the output of a behavior model that describes the behavior of value chain network entity 652). It may include the configuration or conditions of the value chain network entity 652, including data records of the past and current states of the value chain network entity 652. Digital twin 1700 may be continuously updated to reflect current conditions of value chain network entities 652 based on sensor data, test and inspection results, maintenance performed, modifications, etc. Digital twin 1700 may also be configured to communicate with a user through multiple communication channels, such as conversation, text, gestures, etc. For example, digital twin 1700 can receive a query from a user regarding value chain network entity 652, generate a response to the query, and communicate such response to the user. Additionally or alternatively, digital twins 1700 may communicate with each other to learn and identify similar behavioral patterns and problems in other value chain network entities 652, as well as steps to be taken to resolve those problems. You can. Digital twin 1700 can be used for monitoring, diagnosis, simulation, management, remote control, and prognosis, such as optimizing the individual and collective performance and utilization of value chain network entities 652.

For example, machine twin 21010 may continuously capture key operating metrics of machine 724 and may be used to monitor and optimize machine performance in real time. Machine twin 21010 may combine sensor, performance, and environmental data, including insights from similar machines 724, to enable informed maintenance decisions and prediction of the life of various machine components. In embodiments, machine twin 21010 may generate an alarm or other warning based on changes in the operating characteristics of machine 724. The alarm may be due to a problem with a component of machine 724. Additionally, machine twin 21010 can determine similar problems that have previously occurred on the machine or similar machines, provide a description of what caused the problem, what was done to solve the problem, the differences between the current problem and previous problems, and You can explain what actions to take to solve the problem.

Similarly, warehousing twin 1712 may combine a 3D model of the warehouse with inventory and operational data including size, quantity, location, and demand characteristics of different products. Warehousing twin 1712 may also collect data regarding the movement of inventory and personnel within the warehouse, as well as sensor data within connected warehouses. Warehousing twins (1712) can help optimize space utilization and assist in the identification and removal of waste in warehouse operations. Simulation using the warehouse twin 1712 of the movement of product, personnel, and material handling equipment can enable warehouse managers to test and evaluate the potential impact of layout changes or the introduction of new equipment and new processes. .

In embodiments, multiple digital twins of value chain network entities 652 may be integrated, thereby aggregating data across the value chain network to drive entity-level insights as well as system-level insights. For example, consider a simple value chain network with an operating facility 712 containing different machines 724, including conveyors, robots, and inspection devices. Operations facility digital twin 1172 is an overall view of a complete conveyor line within an operations facility 712 (e.g., a warehouse, distribution center, or fulfillment center where packages are moved along the conveyor and inspected before being sent out for delivery). It may be necessary to integrate data from digital twins 1770 of different machines to obtain a picture. While a digital twin of a conveyor line may only provide insight into its performance, a composite digital twin may aggregate data across different machines within an operating facility 712. Accordingly, this can lead to insight into the overall health of the conveyor line within the operating facility 712 by providing an integrated view of individual machines and their interaction with environmental factors within the operating facility. As another example, supply factor twin 1650 and demand factor twin 1640 may be integrated to create an overall picture of the supply-demand equilibrium for product 1510. Integration of the digital twin also enables interrogation of multiple value chain network entities 652 and creates a 360-degree view of the value chain network 668 and its various systems and subsystems.

The ability to integrate digital twins of value chain network entities 652 may be used to create a value chain network digital twin system from multiple digital twin subsystems representing selected entities among supply chain entities, demand management entities, and value chain network entities. It will be clear that it can be done. For example, machine digital twin 1770 is comprised of multiple digital twins of the subsystems and individual components that make up machine 724. A digital twin of a machine can build a model of a machine by integrating all such component twins and their inputs and outputs. Additionally, for example, distribution facility twin system 1714 may be comprised of subsystems such as warehousing twin 1712, order processing twin 1600, and delivery system twin 1610.

Similarly, a process digital twin can be viewed as consisting of a digital twin of multiple subprocesses representing selected entities among supply chain entities, demand management entities, and value chain network entities. For example, a digital twin of a packaging process consists of a digital twin of subprocesses for picking, moving, inspecting, and packing products. As another example, a digital twin of a warehousing process can be viewed as consisting of a digital twin of multiple subprocesses, including receiving, storing, picking, and shipping stored inventory.

It will be clear that a value chain network digital twin system can be created from a plurality of digital twin subsystems, or conversely a digital twin subsystem can be created from a digital twin system, with at least one of the digital twin subsystem and the digital twin system. One represents selected entities among supply chain entities, demand management entities, and value chain network entities.

Similarly, a value chain network digital twin process can be created from a plurality of digital twin subprocesses, or conversely, it can be a digital twin subprocess created from a digital twin process, and at least one of the digital twin subprocess and the digital twin process is a supply. Represents selected entities among chain entities, demand management entities, and value chain network entities.

The analysis obtained from the digital twin 1700 of the value chain network entities 652 and their interactions with each other provides a systematic view of the value chain network as well as its systems, subsystems, processes and subprocesses. This can help generate new insights into how various systems and processes can evolve to improve performance and efficiency.

In embodiments, platform 604 and application 630 may have a system for creating and updating a self-expanding digital twin that represents a set of value chain entities. Self-expanding digital twins continue to learn and expand their scope as they collect more and more data and encounter scenarios. As a result, the self-scaling twin can evolve over time, take on more complex tasks, and answer more complex questions posed by users of the self-scaling digital twin.

In embodiments, platform 604 and application 630 may have a system for scheduling synchronization of change conditions of physical value chain entities to a digital twin representing a set of value chain entities. In embodiments, synchronization between physical value chain entities and their digital twins is on a near real-time basis.

In embodiments, platform 604 and application 630 may extract, share, and/or harmonize data from information technology systems associated with multiple value chain network entities contributing to a single digital twin representing a set of value chain entities. It can have an application programming interface for it.

In an embodiment, the value chain network management platform 604 provides various subsystems that can be implemented as microservices such that other subsystems of the system access the functionality of the subsystem that provides the microservices through an application programming interface API. It can be included. In some embodiments, the various services provided by a subsystem may be deployed as a bundle, integrated by, for example, a set of APIs.

In an embodiment, the value chain network management platform 604 manages the set of value chain network entities for an enterprise and creates, modifies, and manages parameters of digital twins used by the platform to represent the set of value chain network entities. It may include a set of microservices with a set of processing capabilities for at least one of the following:

Value Chain Digital Twin Kit (DTIB)

A value chain network management platform can provide digital twin subsystems in the form of out-of-the-box kit systems with self-configuration capabilities. The kit can provide a data-rich, interactive overview of the set of value chain network entities that make up a subsystem. For example, a supply chain creative digital twin kit system could represent a set of supply chain entities that are associated with the identity of the owner or operator of the supply chain entities. The owner or operator of the supply chain entity can then use the kit to get a holistic picture of the complete portfolio. Owners can research information related to various supply chain entities and ask interactive questions from the digital twin kit system.

In embodiments, a demand management proprietary digital twin kit system may represent a set of demand management entities that are associated with the identity of the owner or operator of the demand management entity.

In an embodiment, a value chain network digital twin kit system to provide unique, self-configuring capabilities includes a set of demand management entities and a set of supply chain entities that are associated with the identity of the owner or operator of the demand management entity and the supply chain entity. can represent.

In embodiments, a warehouse digital twin kit system to provide unique, self-configuring capabilities may represent a set of warehouse entities that are associated with the identity of the owner or operator of the warehouse.

Referring now to Figure 53, an example warehouse digital twin kit system 5000 is shown. Warehouse digital twin kit system 5000 includes a warehousing twin in virtual space 5002 that represents a model of a warehouse 654 in physical space 5004.

Warehouse digital twin kit system 5000 allows an owner or operator 5008 of one or more warehouse entities 654 to obtain a complete portfolio overview of all of these entities, existing or in design or configuration. The owner 5008 navigates through a wealth of information, including warehouse photos 5010, 3D images 5012, live video feeds 5014 of real-time construction progress, and AR or VR renderings 5018 of the warehousing entities 654. can do. Owner 5008 may research the health of one or more entities 654, ask interactive questions, and retrieve detailed information about one or more warehouse entities 654. The warehouse digital twin kit system 5000 can access real-time dynamic data captured by IoT devices and sensors in the warehouse entity 654, interact with the owner 5008, and condition the warehouse entity 654. It can be supported with natural language capabilities that allow you to answer arbitrary questions about .

In an embodiment, the warehouse digital twin kit system 5000 may provide the owner 5008 with an overview of the portfolio of warehouse entities 654 in the form of a 3D information map containing all warehouse entities 654. Owner 5008 can select a specific entity on the map and obtain information about inventory, operational, and health data from warehousing twin 1710. Alternatively, owner 5008 may request information regarding the entire portfolio of owned warehouse entities 654. The warehouse digital twin kit system 5000 integrates information from multiple warehousing twins 1710 and provides a holistic view. The integrated view can help owners 5008 optimize operations across warehouse entities 654 by adjusting inventory locations and staffing levels to match current or expected demand. Owner 5008 may also display information from warehouse digital twin kit system 5000 on a website or marketing materials to be accessed by any customers, suppliers, vendors, and other partners.

In embodiments, a container ship digital twin kit system to provide unique, self-configuring capabilities may represent a set of container ship entities that are associated with the identity of the owner or operator of the container ship.

In embodiments, a port infrastructure digital twin kit system to provide unique, self-configuring capabilities may represent a set of port infrastructure entities that are associated with the identity of the owner or operator of the port infrastructure.

Value Chain Compatibility Testing (VCCT)

Platform 604 uses digital twins 1700 of value chain network entities 652 to test compatibility between different value chain network entities 652 that interact with each other and form various systems and subsystems of the value chain network. It can be placed.

This yields visibility into the compatibility and performance of various systems and subsystems within the value chain network before any physical impact. Any incompatibilities or performance deficiencies in different value chain network entities 652 can be highlighted through digital models and simulations without relying on physical systems to perform such testing, which can be costly and impractical.

The digital twin 1700 may be configured to use an artificial intelligence system 1160 (various expert systems, artificial intelligence systems, neural networks, supervised learning systems, as described throughout this disclosure and in documents incorporated by reference) to perform compatibility testing in the value chain network. , machine learning systems, deep learning systems, and other systems) may be used.

In embodiments, a platform may provide a system for testing the compatibility or configuration of a set of vendor components for a container vessel using a set of digital twins representing the container vessel and vendor components.

In embodiments, a platform may provide a system for testing the compatibility or configuration of a set of vendor components for a warehouse using a set of digital twins representing the warehouse and vendor components.

In embodiments, a platform may provide a system for testing the compatibility or configuration of a set of vendor components for a port infrastructure facility using a set of digital twins representing the port infrastructure facility and the vendor components.

In embodiments, a platform may provide a system for testing the compatibility or configuration of a set of vendor components for a shipyard facility using a set of digital twins representing the shipyard facility and vendor components.

In embodiments, a platform may provide a system for testing the compatibility or configuration of a set of container ships and port infrastructure facilities using a set of digital twins representing the container ships and port infrastructure facilities.

In embodiments, a platform may provide a system for testing the compatibility or configuration of barge and waterway sets for a navigation route using a digital twin set representing the barge and waterway sets.

In embodiments, a platform may provide a system for testing the compatibility or configuration of a set of container ships and cargoes for an identified shipment using a set of digital twins representing the container ships and cargoes.

In embodiments, a platform may provide a system for testing the compatibility or configuration of barge and cargo sets for an identified shipment using a digital twin set representing the barge and cargo.

In embodiments, a platform may provide a system for testing the compatibility or configuration of a set of cargo handling infrastructure facilities and cargo for an identified shipment using a set of digital twins representing the cargo handling infrastructure facilities and cargo. You can.

Value Chain Infrastructure Testing (VCIT)

Platform 604 may deploy digital twin 1700 of value chain network entities 652 to perform stress testing on a set of value chain network entities. Digital twins can help simulate the behavior of value chain network systems and subsystems in a wide variety of environments. Stress testing can help run arbitrary “what-if” scenarios to understand the impact of changes in relevant parameters beyond normal operating values and to assess the resilience of the infrastructure of a value chain network.

Platform 604 uses artificial intelligence systems 1160 (various expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, and others disclosed herein) to perform such stress tests on value chain networks. Learning based on a training set of results, parameters, and data collected from data sources related to a set of value chain network activities to train (including any of the other systems described throughout and in the incorporated document) It may include a system for

In an embodiment, the platform uses a digital twin representing a set of value chain entities to train an artificial intelligence/machine learning system to perform stress tests on machines collected from data sources related to a set of value chain network activities. It may include a system for learning based on a training set of results, parameters, and data.

As explained, the value chain network extends from the acquisition and distribution of raw materials between suppliers and manufacturers, through the delivery, distribution and storage of materials to retailers or wholesalers, and finally to the sale of products to end users. It includes a number of interrelated subsystems and subprocesses that manage and control all aspects associated with the production and delivery of finished products to end users. The complex interconnected nature of value chain networks means that adverse events within one subsystem or one or more value chain entities are reflected throughout the entire value chain network.

Figure 54 is an example method for performing a stress test on a value chain network. Stress testing is intended to test the resilience of a value chain network (including its subsystems) and determine its ability to handle adverse scenarios, such as natural disasters, congested routes, changes in laws, or deep economic recessions. May include simulation exercises. These negative or stressful scenarios may impact one or more entities or subsystems within the value chain network, depending on the nature of the scenario. Therefore, any stress test will require simulating scenarios and analyzing the impact of different scenarios across different subsystems and on the entire value chain network.

At 5102, all historical and current data related to the value chain network is received. The data may include information related to various operating parameters of the value chain network over a specific historical period, for example, the last 12 months. The data can also provide information about typical values of various operating parameters under normal conditions. Some examples of operating parameters include: product demand, procurement turnaround time, productivity, inventory levels in one or more warehouses, inventory turnover, warehousing costs, average time to transport products from a warehouse to a shipping terminal, product Total cost of delivery, service level, etc. At 5104, one or more simulation models of the value chain network are created based on the data. Simulation models help visualize the value chain network as a whole and predict how changes in operating parameters will affect the behavior and performance of the value chain network. In embodiments, the simulation model may be the sum of multiple models of different subsystems of the value chain network.

At 5106, one or more stress scenarios may be simulated by changing one or more parameters beyond normal operating values. Simulation of stress scenarios overcomes the limitations of arbitrary analysis based only on historical data and helps analyze network performance over a range of hypothetical but plausible stress conditions. Simulation involves changing (shocking) one or more parameters while keeping other parameters fixed to analyze the impact of these fluctuations on the value chain network. In embodiments, a single parameter may be changed while keeping the remaining parameters fixed. In other embodiments, multiple parameters may be changed simultaneously. At 5108, the results of stress scenario simulations are determined and the performance of the value chain network and its different subsystems is estimated across various scenarios. At 5110, data, parameters and results are fed to a machine learning process in artificial intelligence system 1160 for further analysis.

The advantage of generating data through simulation and then training machine learning algorithms on this data is the control this approach provides over the volume and frequency of the data, as well as the features in the data.

In an embodiment, the platform collects data from sources related to a set of value chain network activities to train an artificial intelligence/machine learning system to perform stress testing on physical objects using a digital twin representing a set of value chain entities. and a system for learning based on a training set of results, parameters, and data.

In an embodiment, the platform uses a connected network of entities and a digital twin representing a set of value chain entities in the telecommunication network to train an artificial intelligence/machine learning system to perform stress testing on a telecommunication network. It may include a system for learning based on a training set of results, parameters, and data collected from data sources related to the set of activities.

For example, a telecommunication network may be stress tested for resilience by intentionally increasing network traffic by generating data packets and sending them to specific target nodes within the telecommunication network. Additionally, the amount of traffic can be varied to create variable load conditions on the target node by manipulating the number, rate or amount of data within the data packets. A response from the target node can be determined to evaluate how the node performed in the stress test. Target nodes may be selected from different parts of the telecommunication network for stress testing to test the robustness of any part of the network in any topology. Simulated stress testing for telecommunications networks can be used to identify vulnerabilities in any part of the network, so that vulnerabilities can be remedied before users experience network outages in deployed networks.

In embodiments, a platform may include a system for using a digital twin to represent a set of value chain entities in a demand management environment to perform a set of stress tests on a set of workflows in the demand management environment using the digital twin. Stress testing can demonstrate the impact on a digital twin of changing a set of demand-related parameters to levels that exceed normal operating levels. For example, the demand for a product in a value chain network can be affected by factors such as changes in consumer confidence, recession, excessive inventory levels, replacement product prices, overall market index, exchange rate changes, etc. Demand factor twin 1640 can simulate such scenarios by varying supply parameters and evaluate the impact of such stresses on the demand environment 672. Stress testing performed using digital twins can help test and evaluate the resilience of value chain networks in both over- and under-demand cases.

In embodiments, a platform may include a system for using a digital twin representing a set of value chain entities in a supply chain to perform a set of stress tests on a set of workflows in the supply chain using the digital twin. Stress testing demonstrates the impact on the digital twin of changing a set of supply chain-related parameters to levels that exceed normal operating levels. For example, the supply of products in a value chain network can be affected by factors such as weather, natural disasters, traffic congestion, changes in regulations including taxes and subsidies, and border restrictions. Supply factor twin 1650 can simulate such scenarios by varying supply parameters and evaluate the impact of such stresses on supply environment 670. Stress testing performed using digital twins can help test and evaluate the resilience of value chain networks in both oversupply and undersupply events.

Value Chain Incident Management (VCIM)

Platform 604 may deploy digital twin 1700 of value chain network entities 652 to automatically manage a set of events related to a set of value chain network entities and activities. Events may include, among other things, accidents, fires, explosions, labor strikes, increases in tariffs, changes in the law, changes in market prices (e.g. of fuels, components, materials, or end products), changes in demand, and the activities of cartels. , closure of boundaries or routes, and/or any event causing disruption to the value chain network, such as natural events and/or disasters (including storms, heat waves, winds, earthquakes, floods, hurricanes, tsunamis, etc.). It can be included.

Additionally, platform 604 may provide real-time visualization and analysis of mobility flows in the value chain network. This can help quantify risk, improve visibility and react to disruptions in the value chain network. For example, real-time visualization of utility flows for delivery activities using a digital twin can help detect the occurrence and location of an emergency involving the delivery system and deploy emergency services to the detected location.

In embodiments, the platform may deploy a digital twin 1700 of a value chain network entity 652 for more accurate determination of fault in an accident. The platform includes a monitoring layer 614 and a data storage layer 624 to train an artificial intelligence system 1160 using a set of digital twins 1700 of the involved value chain network entities 652 to determine fault in an incident. ) can learn based on a training set of accident results, parameters, and data collected from data sources. For example, data from the digital twins of two crashing vehicles, in addition to data from drivers, witnesses and police reports, could be compared to each other to determine fault in an accident.

In an embodiment, the platform may be used to train an artificial intelligence system 1160 to use the digital twin 1700 of a selected set of value chain network entities 652 to detect occurrences of fraud. The system may include a system for learning based on a training set of vehicle event results, parameters, and data collected from data sources associated with the set. For example, comparing vehicle event data from a vehicle's digital twin to any insurance claims, contract claims, or maritime claims for that vehicle can help detect any mismatches between the two.

In an embodiment, the platform may use an artificial intelligence system (1700) of a selected set of value chain network entities 652 to detect unreported abnormal events for the selected set of value chain network entities 652. and a system for learning based on a training set of vehicle results, parameters, and data collected from data sources associated with a set of value chain network entities 652 to train 1160). Consider an example where a vehicle's digital twin indicates an unusual event, such as an accident, but this event was not reported by the vehicle's driver. Unreported events may be added to the vehicle and driver records by the vehicle's lessor. Additionally, the lessor of the vehicle may charge the lessee for any repairs or diminished value of the vehicle at the end of the lease and adjust its estimated residual value accordingly. Similarly, underwriters can add unreported events to vehicle and driver records. Reports may detail the exact nature of the incident, timing, location, defects, etc., or simply the fact that there was an unreported incident. This information can then be used to calculate insurance premiums.

Finally, if there are multiple entities involved in an incident, the data can be triangulated with digital twins of other entities for verification.

Value Chain Predictive Maintenance (PMVC)

Platform 604 may deploy digital twin 1700 of value chain network entities 652 to predict when a set of value chain network entities may require maintenance.

By examining historical and current operational data, a digital twin can predict the expected wear and failure of a system's components, thereby reducing the risk of unplanned downtime and the need for scheduled maintenance. Instead of over-servicing or over-maintaining products to avoid costly downtime, repairs, or replacements, any product performance issues predicted by the digital twin are resolved proactively or just-in-time. It can be.

Digital twin 1700 may collect event or state data about value chain entities 652 from monitoring layer 614 and historical or other data from selected data sources in data storage layer 624. Predictive analytics powered by artificial intelligence system 1160 analyzes data, retrieves correlations, and generates predictions regarding maintenance needs and remaining useful life of a set of value chain entities 652.

Platform 604 includes artificial intelligence systems 1160 (a variety of expert systems, artificial intelligence systems, neural networks, supervised learning, etc.) to perform condition monitoring, anomaly detection, failure prediction, and predictive maintenance of a set of value chain entities 652. Results collected from data sources related to a set of value chain network activities to train systems, machine learning systems, deep learning systems, and any of the other systems described throughout this disclosure and in incorporated documents) , parameters, and a system for learning based on a training set of data.

In embodiments, the platform uses the machine's digital twin to collect data, parameters, and machine maintenance from data sources related to a set of machine activities to train an artificial intelligence/machine learning system to perform predictive maintenance on the machine. It may include a system for learning based on a training set of reward results.

In embodiments, artificial intelligence system 1160 may train models, such as predictive models (e.g., various types of neural networks, classification-based models, regression-based models, and other machine learning models). In embodiments, training may be supervised, semi-supervised, or unsupervised. In embodiments, training may be performed using training data that may be collected or generated for training purposes.

Example artificial intelligence system 1160 trains a machine predictive maintenance model. A predictive maintenance model may be a model that receives machine-related data and outputs one or more predictions or answers regarding the remaining life of the machine. Training data may be collected from a number of sources, including machine specifications, environmental data, sensor data, execution information, results data, and notes maintained by machine operators. Artificial intelligence system 1160 takes raw data, preprocesses it, and applies machine learning algorithms to create a predictive maintenance model. In an embodiment, artificial intelligence system 1160 may store the predictive model in a model datastore within data storage layer 624.

Some examples of questions that predictive models can answer are: when will the machine fail, what type of failure will it be, what is the probability that the failure will occur in the next Are the machines behaving in an uncharacteristic way? Which machines are in most urgent need of maintenance?

Artificial intelligence system 1160 may train multiple predictive models to answer different questions. For example, a classification model can be trained to predict failure within a given time window, while a regression model can be trained to predict the remaining useful life of a machine.

In embodiments, training may be performed based on feedback received by the system, also referred to as “reinforcement learning.” In embodiments, artificial intelligence system 1160 may receive a set of circumstances (e.g., properties of the machine, properties of the model, etc.) that led to the prediction and results associated with the machine and update the model according to the feedback. there is.

In embodiments, artificial intelligence system 1160 may use clustering algorithms to identify failure patterns hidden in failure data to train models for detecting non-characteristic or abnormal behavior. Failure data and historical records across multiple machines can be clustered to understand how different patterns correlate with specific wear behavior and develop maintenance plans that resonate with the failures.

In an embodiment, artificial intelligence system 1160 may output a score for each possible prediction, where each prediction corresponds to a possible outcome. For example, when using a predictive model that is used to determine the likelihood that a machine will break down in the next week, the predictive model might output a score for the “will break” outcome and a score for the “won’t break” outcome. there is. Artificial intelligence system 1160 can then select the outcome with the greater score as the prediction. Alternatively, system 1160 may output each score to the requesting system. In an embodiment, the output from system 1160 includes a probability of accuracy of the prediction.

55 is an example method used by machine twin 1770 to detect faults and predict any future failure of machine 724.

At 5202, multiple streams of machine-related data from multiple data sources are received at machine twin 1770. This includes machine specifications such as mechanical characteristics, data from maintenance records, operational data collected from sensors, historical data including failure data from multiple machines running at different times and under different operating conditions, etc. At 5205, the raw data is cleaned by removing any missing or noisy data that may have occurred due to any technical problems in the machine during collection of the data. At 5208, one or more models are selected for training by machine twin 1770. Selection of a model is based on the type of data available in the machine twin 1770 and the desired results of the model. For example, there may be cases where failure data from a machine is not available or only a limited number of failure data sets exist because routine maintenance is performed. Classification or regression models may not work well for such cases and clustering models may be best. As another example, if the desired outcome of the model is to determine the current condition of the machine and detect any defects, a defect detection model may be selected, whereas if the desired outcome of the model is to predict future failures, a remaining useful life prediction model may be selected. This can be selected. At 5210, one or more models are trained using a training data set and tested for performance using a test data set. At 5212, the trained model is used to detect defects and predict future failures of the machine against production data.

56 is an example embodiment showing the deployment of machine twin 21010 to perform predictive maintenance on machine 724. Machine twin 1770 receives data from data storage system 624 on a real-time or near real-time basis. Data storage system 624 may store different types of data in different data stores. For example, machine datastore 5202 may store data related to machine identification and attributes, machine status and event data, data from maintenance records, historical operating data, notes from machine operators, etc. Sensor datastore 5204 may store sensor data from motion, such as temperature, pressure, and vibration, which may be stored as signals or time series data. Failure datastore 5310 may store failure data from machine 724 or similar machines running at different times and under different operating conditions. Model datastore 5312 may store data related to different prediction models, including defect detection and remaining life prediction models.

Machine twin 1770 then coordinates with the artificial intelligence system to select one or more of the models based on the type and quality of available data and the desired answer or outcome. For example, if the intended use of machine twin 1770 is to simulate hypothetical scenarios and predict how the machine will behave under those scenarios, physical model 5320 may be selected. A fault detection and diagnosis model 5322 may be selected to determine the current health and any fault conditions of the machine. A simple fault detection model may use one or more condition indicators to distinguish between regular and defective behavior and may have thresholds for the condition indicators that, when exceeded, indicate a fault condition. More complex models may train a classifier to compare the values of one or more condition indicators with values associated with a failure state and return the probability of the presence of one or more failure states.

The Remaining Useful Life (RUL) prediction model 5324 is used to predict future failure and may include a degradation model 5326, a survival model 5328, and a similarity model 5330. An example RUL prediction model may fit the time evolution of a condition indicator and predict how long it will be before the condition indicator crosses some threshold indicating a failure. Other models can compare the time evolution of condition indicators with measured or simulated time series from similar systems that ran until failure.

In embodiments, a combination of one or more of these models may be selected by machine twin 1770.

The artificial intelligence system 1160 may include a machine learning process 5340, a clustering process 5342, an analysis process 5344, and a natural language process 5348. Machine learning process 5340 operates with machine twin 1770 to train one or more models as identified above. An example of such a machine learning model is RUL prediction model 5324. Model 5324 can be trained using training data set 5350 from data storage system 624. The performance of model 5324 and classifier can then be tested using test data set 5350.

Clustering process 5342 may be implemented to identify failure patterns hidden in the failure data to train a model for detecting non-characteristic or abnormal behavior. Failure data and historical records across multiple machines can be clustered to understand how different patterns correlate with specific wear behavior. Analysis process 5344 performs data analysis on various data to identify insights and predict outcomes. Natural language process 4348 coordinates with machine twin 1770 to communicate results and results to users of machine twin 1770.

Results 5360 may be in the form of modeling results 5362, alerts and warnings 5364, or remaining useful life (RUL) predictions 5368. Machine twin 1770 may communicate with the user through multiple communication channels such as conversation, text, and gestures to convey results 5360.

In embodiments, the model may then be updated or enhanced based on model results 5360. For example, an artificial intelligence system can receive a set of circumstances that led to a failure and a prediction of the outcome, and update the model based on the feedback.

In embodiments, the platform uses a digital twin of a vessel to train an artificial intelligence/machine learning system to perform predictive maintenance on a vessel, a training set of vessel maintenance results collected from data sources related to a set of vessel activities. , parameters, and a system for learning based on data.

In embodiments, the platform uses a digital twin of a barge to train an artificial intelligence/machine learning system to perform predictive maintenance on a barge, including a training set of barge maintenance results collected from data sources associated with a set of barge activities; It may include a system for learning parameters and data.

In an embodiment, the platform uses a digital twin of a port infrastructure facility to train an artificial intelligence/machine learning system to perform predictive maintenance on the port infrastructure facility, collected from data sources related to a set of port activities. It may include a system for learning based on a training set of maintenance results, parameters, and data.

In embodiments, the platform collects data from data sources associated with a set of value chain entities to train an artificial intelligence/machine learning system to use a digital twin of a selected set of value chain entities to estimate the cost of repair of a damaged object. , a system for learning based on a training set of parameters, and mathematical results.

In embodiments, the platform uses a digital twin of the infrastructure to train an artificial intelligence/machine learning system to predict infrastructure degradation based on a training set of infrastructure results, parameters, and data collected from data sources. May include a system for learning.

In an embodiment, the platform uses a digital twin of a city to train an artificial intelligence/machine learning system to model natural hazard risks for a set of delivery infrastructure facilities, using natural hazards collected from data sources associated with a set of delivery activities. It may include a system for learning based on a training set of results, parameters, and data.

In embodiments, the platform uses a digital twin of the set of facilities to train an artificial intelligence/machine learning system to monitor delivery infrastructure maintenance activities for a set of delivery infrastructure facilities using data sources related to the set of delivery activities. and a system for learning based on a training set of maintenance results, parameters, and data collected from.

In embodiments, the platform uses a digital twin of a set of delivery infrastructure facilities to train an artificial intelligence/machine learning system to detect the occurrence and location of maintenance issues, collected from data sources related to a set of delivery activities. It may include a system for learning based on a training set of maintenance results, parameters, and data, and has a system for automatically deploying maintenance services to detected locations.

Referring to FIG. 57 , platform 604 includes, integrates with, manages, controls, coordinates, or otherwise integrates customer digital twin 5502 and/or customer profile digital twin 1730. It can be treated in this way.

Customer digital twin 5502 may represent an evolving and continuously updated digital representation of a value chain network customer 662. In embodiments, Value Chain Network Customers 662 may include consumers, licensees, operators, enterprises, value-added resellers and other resellers, distributors, and retailers (online retailers, mobile retailers, traditional brick-and-mortar retailers, pop-up retailers). (including stores, etc.), end users, and others who may purchase, license, or otherwise use categories of goods and/or related services.

On the other hand, the customer profile digital twin 1730 may provide one or more demographic information (age, gender, race, marital status, number of children, occupation, annual income, education level, living status (homeowner, renter, etc.)) of a set of customers. , psychological, behavioral, economic, geographic, physical (e.g., size, weight, health status, physiological state or condition, etc.), or other attributes. In embodiments, the customer profile digital twin 1730 may be an enterprise customer profile digital twin that represents the attributes of a set of enterprise customers. In embodiments, the customer profiling application may include historical purchase data, loyalty program data, behavior tracking data (customer interaction with intelligent product 1510), and may be used to manage customer profiles 5504 based on data (including data captured from interactions), online clickstream data, interactions with intelligent agents, and other data sources.

Customer 662 may, for example, convert customer digital twin 1730 into a value chain network data object 1004, such as event data 1034, state data 1140, or other data for value chain network customer 662. By populating, it can be represented as a set of one or more customer digital twins 5502. Likewise, customer profile 5504 can be formed into one or more customer profile digital twins 1730, e.g., by populating customer profile digital twins 1730 with value chain network data objects 1004, as described throughout this disclosure. It can be displayed as a set of .

Customer digital twin 5502 and customer profile digital twin 1730 may allow for modeling, simulation, prediction, decision making, classification, etc.

If customers 662 are consumers, for example, each customer digital twin 1730 may include identity data, account data, payment data, contact data, age data, gender data, race data, location data, demographic data, Life status data, mood data, stress data, behavior data, personality data, interest data, preference data, style data, medical data, physiological data, psychological data, physical attribute data, education data, employment data, salary data, net worth data , family data, household data, relationship data, pet data, contact/connection data (e.g. mobile phone contacts, social media connections, etc.), transaction history data, political data, travel data, product interaction data, product feedback data, Customer service interaction data (e.g., communication with a chatbot, or telephone communication with a customer service agent in a call center), fitness data, sleep data, nutritional data, software program interaction observation data 1500 (e.g., by customers interacting with various software interfaces of applications 630 accompanying value chain entities 652) and physical process interaction observation data 1510 (e.g., with products or other value chain entities 652). (by observing customers interacting), etc.

In another example where customer 662 is a business or business, customer digital twin 1730 may include identity data, account data, payment data, transaction data, product feedback data, location data, revenue data, business type data, products and/or It may be populated with service offer data, worker data (such as identity data, role data, etc.), and other enterprise-related attributes.

Customer Digital Twin and Customer Profile Digital Twin 1730 is used for visualization of value chain network customers 662 and customer profiles 5504, as well as coordinated intelligence (artificial intelligence system 1160) enabled or facilitated by the digital twin. ), edge intelligence, analytics, and other capabilities) and other value-added services and capabilities.

In an embodiment, customer digital twin 5502 and customer profile digital twin 1730 may take advantage of the presence of multiple applications 630 within value chain management platform 604, such that a pair of applications can be used to manage value chain entities 652 ) and other inputs (e.g., from the monitoring layer 614), as well as data sources collected for (e.g., from the data storage layer 624), as well as (referenced throughout and by reference throughout this disclosure) Through the use of artificial intelligence systems 1160 (including any of the various expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, and other systems described in the incorporated documents) and Events, state information, and outputs may be shared, including through the use of content collected by the monitoring layer 614 and data collection system 640, which collectively provide much needed information to enrich content in the digital twin. It can provide a richer environment.

An environment for the development of a customer digital twin 5502 allows developers to take input from selected data sources in the data storage layer 624 and events or other data from the monitoring system layer 614 and include them in the customer digital twin 5502. It may include a set of interfaces for developers to configure the artificial intelligence system 1160 to provide them to do so. A customer digital twin development environment may be configured to take output and results from various applications 630. In embodiments, customer digital twin 1730 may be provided for a wide range of value chain network applications 630 mentioned throughout this disclosure and documents incorporated by reference herein.

In embodiments, customer digital twin 5502 may be rendered by a computing device so that a user can view a digital representation of customer 714. For example, customer digital twin 5502 may be rendered and output to a display device. In another example, 5502 may be rendered in a three-dimensional environment and viewed using a virtual reality headset.

The environment for development of the customer profile digital twin 1730 may include a set of interfaces for developers to receive input from selected data sources in the data storage layer 624 and events from the monitoring system layer 614. Alternatively, the artificial intelligence system 1160 can be configured to take other data and feed it for inclusion in the customer profile digital twin 1730. The customer profile digital twin development environment can be configured to take output and results from various applications 630. In embodiments, customer profile digital twin 1730 may be provided for a wide range of value chain network applications 630 referenced throughout this disclosure and documents incorporated by reference herein.

In embodiments, adaptive intelligence system layer 614 is configured to train and implement artificial intelligence systems to perform tasks related to value chain network 668 and/or value chain network entities 652. For example, adaptive intelligence system layer 614 may recommend products, enhance customer experience, select advertising attributes for advertisements related to value chain products and/or services, and/or other appropriate value chain tasks. It can be used for.

In embodiments, customer profile digital twin 1730 or another customer digital twin may, for example, allow a customer to request, select, or modify a set of characteristics, states, behaviors, or other aspects represented in digital twin 1730. It can be created interactively and collaboratively with customers, allowing them to delete, delete, or otherwise influence it. For example, a customer may improve size (e.g., shoe size, dress size, shirt size, pant size, etc.), indicate interests and needs (e.g., what the customer is interested in purchasing), It can display behavior (e.g., a project planned by an enterprise), update current status (e.g., to reflect changes), etc. Accordingly, a version of the digital twin 1730 may be made available to the customer, such as in a graphical user interface, where the customer can manipulate one or more aspects of the digital twin 1730, request changes, etc. there is. In embodiments, there may be a version for customer review, an internal version for the business or host, a version for each of the brands in a particular set (e.g., if a customer's appropriate clothing size varies across brands), and a public version (e.g., feedback from friends). There are multiple versions of the digital twin 1730, such as a personal version (e.g., shared to the customer's social networks), a simulated version, a real-time version, etc. Can be maintained for a given customer. In embodiments, adaptive intelligence system layer 614 is configured to utilize customer digital twins 5502, customer profile digital twins 1730, and/or other digital twins 1700 of other value chain network entities 652. do. In embodiments, adaptive intelligence system layer 614 is configured to perform simulations using customer digital twin 5502, customer profile digital twin 1730, and/or digital twins of other value chain network entities 652. do. For example, adaptive intelligence system layer 614 may change one or more characteristics of product digital twin 1780 when its use is simulated by customer digital twin 1730.

In embodiments, simulation management system 5704 may set up, provision, configure, and otherwise manage simulations and interactions between digital twins 1700 representing value chain entities 652.

In an embodiment, adaptive intelligence system layer 614 may, for each set of features, run a simulation based on the set of features and collect simulation result data resulting from the simulation. For example, in running a simulation involving the interaction of an intelligent product digital twin 1780 representing an intelligent product 1510 and a customer digital twin 1730, the adaptive intelligence system layer 614 may The dimensions of 1780 can be varied and simulations can be run to produce results in simulation management system 5704. In this example, the result may be the amount of time taken by the customer digital twin 5502 to complete a task using the intelligent product digital twin 1780. During the simulation, the adaptive intelligence system layer 614 determines the intelligent product digital twin 1780 display screen size, available capabilities (processing, conversation recognition, speech recognition, touch interface, remote control, self-organization, self-healing, process automation, computation, artificial intelligence, data storage, etc.), materials, and/or any other characteristics of the intelligent product digital twin 1780. Simulation data 5710 may be generated for each simulation and may include result data as well as feature data used to perform the simulation. In the example described above, simulation data 5710 may be the characteristics and results resulting therefrom of customer digital twin 5502 and intelligent product digital twin 1780 used to perform the simulation. In embodiments, machine learning system 5720 may use training data 5730, results data 5740, simulation data 5710, and/or data from other types of external data sources 5702 (weather data, stock market data, sports event data, news event data, etc.) can be received. In embodiments, this data may be provided to machine learning system 5720 via an API in adaptive intelligence system layer 614. Machine learning system 5720 may use the received data (training data, results data, simulation data, etc.) to train, retrain, or enhance machine learning model 5750.

58 illustrates an example of an advertising application interfacing with the adaptive intelligence system layer 614. In an example embodiment, an advertising application may be configured to automate advertising-related tasks for value chain products or services.

In embodiments, machine learning system 5720 may be one or more utilized by artificial intelligence system 1160 to make classifications, predictions, and/or other decisions related to advertising for a set of value chain products and/or services. Train model 5750.

In an example embodiment, model 5750 is trained to select advertising features to optimize one or more outcomes (e.g., to maximize product sales for product 1510 in value chain network 668). . Machine learning system 5720 can train model 5750 using n-tuples containing features associated with an advertisement and one or more results associated with the advertisement. In this example, the characteristics for the advertisement include the product and/or service category advertised, the product features advertised (price, product vendor, etc.), the service features advertised, and the type of advertisement (television, radio, podcast, social media, email, etc.). ), ad length (10 seconds, 30 seconds, etc.), ad timing (morning, pre-holiday, etc.), ad tone (comedy, informational, emotional, etc.), and/or other relevant ad characteristics. It doesn't work. In this example, results related to an advertisement may include product sales, total cost of the advertisement, advertisement interaction measures, etc. In this example, one or more digital twins 1700 may be used to simulate different arrangements (e.g., digital twins of advertisements, customers, customer profiles, and environments), whereby one or more characteristics of the digital twins may be used to simulate different arrangements. Changes can be made for each simulation, and the results of each simulation can be recorded in a tuple with appropriateness. Other examples of training advertising models may include models being trained to generate advertisements for value chain products 650, models being trained to manage advertising campaigns for value chain products 650, etc. In operation, artificial intelligence system 1160 may use such model 5750 to make advertising decisions on behalf of advertising application 5602 given one or more characteristics associated with an advertising-related task or event. For example, artificial intelligence system 1160 may select a type of advertisement (e.g., social media, podcast, etc.) to use for value chain product 1510. In this example, advertising application 5602 may provide product features to artificial intelligence system 1160. These characteristics may include the product vendor, the price of the product, etc. In embodiments, artificial intelligence system 1160 may insert these features into one or more of models 5750 to obtain one or more decisions, which may include which type of advertisement to use. In embodiments, artificial intelligence system 1160 may utilize customer digital twin 5502 and/or customer profile digital twin 1730 to run a simulation of one or more decisions and generate simulation data 5710. Machine learning system 5720 may receive simulation data 5710 and other data as described throughout this disclosure to retrain or strengthen machine learning models. In an embodiment, customer digital twin 5502, customer profile digital twin 1730, and other digital twins 1700 process decisions made by artificial intelligence system 1160 before providing decisions to value chain entity 652. It may be utilized by the artificial intelligence system 1160 to simulate. In this example, customer profile digital twin 1730 may be utilized by artificial intelligence system 1160 to simulate decisions made by artificial intelligence system 1160 before providing decisions to advertising application 5602. In embodiments where simulation results are unacceptable, simulation data 5710 may be reported to machine learning system 5720, which may use the received data to retrain machine learning model 5750. and the machine learning model can then be utilized by the artificial intelligence system 1160 to make new decisions. Advertising application 824 may use the decision(s) made by artificial intelligence system 1160 to initiate an advertising event. In embodiments, following an advertising event, the outcome of the event (e.g., product sales) may be reported to machine learning system 5720 to strengthen model 5750 used to make decisions. Additionally, in some embodiments, the output of the advertising application and/or other value chain entity data source updates one or more characteristics of customer digital twin 5502, customer profile digital twin 1730, and/or other digital twin 1700. can be used to

59 illustrates an example of an e-commerce application 5604 integrated with an adaptive intelligence system layer 614. In an embodiment, e-commerce application 5604 may be configured to generate product recommendations for value chain customers 662. For example, e-commerce application 5604 may be configured to receive one or more product characteristics for value chain network product 1510. Examples of product characteristics may include, but are not limited to, product type, product capabilities, product price, product materials, product vendor, etc. In an embodiment, e-commerce application 5604 determines recommendations to optimize results. Examples of results include observations of software interactions (e.g., mouse movements, mouse clicks, cursor movements, navigation, etc.), such as those logged and/or tracked by the software interaction observation system 1500, purchases of products by customer 714, etc. actions, menu selection, etc.). In embodiments, e-commerce application 5604 may interface with artificial intelligence system 1160 to provide product features and receive product recommendations based thereon. In embodiments, artificial intelligence system 1160 may utilize one or more machine learning models 5750 to determine recommendations. In some embodiments, simulations run by customer digital twin 1730 may be used to train a product recommendation machine learning model.

60 is a schematic diagram illustrating an example of a demand management application 824 integrated with an adaptive intelligence system layer 614. In an embodiment, the artificial intelligence system 1160 uses a machine learning model 5750 trained to make demand management decisions for the demand environment 672 on behalf of the demand management application 824, given one or more demand factors 644. ) can be used. Demand factors 644 may include product type, product capability, product price, product material, time of year, location, etc. In embodiments, artificial intelligence system 1160 may determine demand management decisions for value chain product 1510. For example, artificial intelligence system 1160 may generate a demand management decision regarding how many printer ink cartridges should be supplied to a particular area during an upcoming month. In this example, demand management system 824 may provide demand factors 644 to artificial intelligence system 1160. In embodiments, artificial intelligence system 1160 may insert these factors 644 into one or more machine learning models 5750 to obtain one or more demand management decisions. These decisions may include the volume of ink cartridges that must be transferred to the selected area during the selected month.

In embodiments, artificial intelligence system 1160 may utilize customer profile digital twin 1730 to run simulations for proposed decisions related to demand management. Demand management application 824 may then use the decision(s) made by artificial intelligence system 1160 to initiate an ink resupply event. Additionally, following an ink refill event, the outcome of the event (e.g., ink cartridge sold) may be reported to machine learning system 5720 to strengthen the model used to make decisions. Additionally, in some embodiments, the output of demand management system 824 and/or other value chain entity data sources may be used to update one or more characteristics of customer profile digital twin 1730 and/or other digital twin 1700. You can.

In embodiments, the API allows users to access customer digital twin 5502 and/or customer profile digital twin 1730. In embodiments, the API allows a user to receive one or more reports related to a digital twin.

Platform 604 may include, integrate, be integrated with, manage, control, coordinate, or otherwise handle home demand digital twin 5902. Household demand digital twin 5902 may be a digital representation of household demand for a product category or set of product categories.

The environment for development of the household demand digital twin 5902 may include a set of interfaces for developers, where the developer may receive input from selected data sources in the data storage layer 624 and input from the monitoring system layer 614. Artificial intelligence system 1160 can be configured to take events or other data and feed them for inclusion in household demand digital twin 5902. The home demand digital twin development environment can be configured to take output and results from various applications 630. In embodiments, household demand digital twin 5902 may be provided for a wide range of value chain network applications 630 referenced throughout this disclosure and documents incorporated by reference herein.

In embodiments, digital twin 1700 may be created from another digital twin. For example, customer digital twin 5502 can be used to create an anonymized customer digital twin 5902. The platform may include, integrate, be integrated with, manage, control, coordinate, or otherwise handle the anonymized customer digital twin 5902. Anonymized customer digital twin 5902 may be an anonymized digital representation of customer 714. In embodiments, anonymized customer digital twin 5902 may not be populated with personally identifiable information, but may be populated using the same data source as the otherwise corresponding customer digital twin 5502.

In an embodiment, an environment for the development of anonymized customer digital twins 1730 allows developers to take inputs from selected data sources in the data storage layer 624 and events or other data from the monitoring system layer 614 and anonymize them. It may include a set of interfaces for developers to configure the artificial intelligence system 1160 to supply them for inclusion in the customer digital twin 5902. The anonymized digital twin development environment can be configured to take output and results from various applications 630. In embodiments, anonymized customer digital twin 5902 may be provided for a wide range of value chain network applications 630 referenced throughout this disclosure and documents incorporated by reference herein.

In an embodiment, anonymized customer digital twin 5902 includes an API that can receive requests for access to anonymized customer digital twin 5902. The requesting entity may issue an access request using the API of the anonymized customer digital twin 5902. Access requests can be routed from the API to the access logic of the anonymized customer twin 5902, which can determine whether the requesting entity is entitled to access. In embodiments, users may monetize access to the anonymized customer digital twin 5902, such as by subscription or any other suitable monetization method.

Platform 604 may include, integrate, be integrated with, manage, control, coordinate, or otherwise handle an enterprise customer engagement digital twin. An enterprise customer engagement digital twin may be a digital representation of a set of attributes of an enterprise customer related to the customer's engagement with the enterprise's set of offerings.

The environment for the development of an enterprise customer engagement digital twin allows developers to take inputs from selected data sources in the data storage layer 624 and events or other data from the monitoring system layer 614 and include them in the enterprise customer engagement digital twin. It may include a set of interfaces for developers to configure the artificial intelligence system 1160 to provide The enterprise customer engagement digital twin development environment can be configured to take output and results from various applications 630. In embodiments, enterprise customer engagement digital twins may be provided for the broad range of value chain network applications 630 referenced throughout this disclosure and documents incorporated by reference herein.

Referring to FIG. 61 , platform 604 may include, integrate, be integrated with, manage, control, coordinate, or otherwise handle component digital twin 6002. Component digital twin 6002 may represent an evolving, continuously updated digital profile of component 6002 of value chain product 650. Component digital twin 6002 may allow for modeling, simulation, prediction, decision making, classification, etc.

A product component can be created into one, for example, by populating the component digital twin 6002 with value chain network data objects 1004, such as event data 1034, state data 1140, or other data for the value chain network product component. Or it may be displayed as a set of component digital twins 6002.

Product 1510 includes finished products, software products, hardware products, component products, materials, equipment items, consumer packaging products, consumer products, food and beverage products, household products, business supply products, consumable products, pharmaceutical products, medical device products, It may be a product of any category, such as a technology product, an entertainment product, or a set of any other type of product and/or related service, and may in embodiments include, but are not limited to, data processing, networking, sensing, autonomous operation, Intelligent agents, natural language processing, conversation recognition, speech recognition, touch interfaces, remote control, self-organization, self-healing, process automation, computation, artificial intelligence, analog or digital sensors, cameras, sound processing systems, data storage, data integration. , and/or intelligent products 1510 enabled with a set of capabilities, such as various Internet of Things capabilities. Component 6002 may be any category of product component.

As an example, component digital twin 6002 may be populated with supplier data, dimensional data, material data, thermal data, pricing data, etc.

Component Digital Twin 6002 provides value-added services and capabilities, as well as coordinated intelligence (including artificial intelligence systems 1160, edge intelligence, analytics, and other capabilities) that are enabled or facilitated by the Component Digital Twin. It may include a set of components, processes, services, interfaces, and other elements for the development and deployment of digital twin capabilities for visualization of chain network components 714.

In embodiments, component digital twin 6002 may take advantage of the presence of multiple applications 630 within value chain management platform 604, such that a pair of applications can be used to generate data sources ( e.g., from the data storage layer 624) and other inputs (e.g., from the monitoring layer 614), but may also be used by artificial intelligence systems 1160 (various expert systems, artificial intelligence systems, neural networks, through the use of supervised learning systems, machine learning systems, deep learning systems, and other systems described throughout this disclosure and in documents incorporated by reference) and monitoring layer 614 and data collection. Outputs, events, state information, and outputs can be shared, including through the use of content collected by system 640, which together form an even richer environment for enriching content in component digital twin 6002. It can be provided.

An environment for the development of component digital twins 6002 allows developers to take inputs from selected data sources in the data storage layer 624 and events or other data from the monitoring system layer 614 and compile them into component digital twins 6002. It may include a set of interfaces for developers to configure the artificial intelligence system 1160 to supply for inclusion. The component digital twin development environment can be configured to take output and results from various applications 630. In embodiments, component digital twin 6002 may be provided for a wide range of value chain network applications 630 mentioned throughout this disclosure and documents incorporated by reference herein. In embodiments, digital twin 650 may be created from another digital twin 1700. For example, product digital twin 1780 may be used to create component digital twin 6002. In another example, component digital twin 6002 can be used to create product digital twin 1780. In embodiments, digital twin 1700 may be embedded in another digital twin 1700. For example, component digital twin 6002 may be embedded in product digital twin 1780, which may be embedded in environmental digital twin 6004.

In embodiments, simulation management system 6110 may set up, provision, configure, and otherwise manage simulations and interactions between digital twins 1700 representing value chain entities 652.

In embodiments, adaptive intelligence system layer 614 may use component digital twins 6002 and/or digital twins 1700 of other value chain network entities 652 to run simulations in simulation management system 6110. It is composed. For example, adaptive intelligence system layer 614 may adjust one or more characteristics of environmental digital twin 6004 as the set of component digital twins 6002 experiences the environment. In an embodiment, adaptive intelligence system layer 614 may, for each set of features, run a simulation based on the set of features and collect simulation result data resulting from the simulation.

For example, in running a simulation on a set of component digital twins 6002 representing components of a value chain product 1510 in environmental digital twin 6004, adaptive intelligence system layer 614 may The characteristics of (6110) can be varied and simulations can be run to produce results. During simulation, adaptive intelligence system layer 614 may vary environmental digital twin temperature, pressure, lighting, and/or any other characteristics of environmental digital twin 6004. In this example, the result may be the condition of component digital twin 6002 after high temperature has been applied. Results from the simulation can be used to train a machine learning model 6120.

In embodiments, machine learning system 6150 may use training data 6170, results data 6160, simulation data 6140, and/or data from other types of external data sources 6150 (weather data, stock market data, sports event data, news event data, etc.) can be received. In embodiments, this data may be provided to machine learning system 6150 via an API in adaptive intelligence system layer 614. In an embodiment, machine learning system 6150 may receive simulation data 6140 related to a simulation of component digital twin 6002. In this example, simulation data 6140 may be the characteristics of the component digital twin 6002 used to perform the simulation and the results resulting therefrom.

In embodiments, machine learning system 6150 may train/strengthen machine learning model 6120 using received data to improve the model.

62 illustrates an example of a risk management system 6102 interfacing with an adaptive intelligence system layer 614. In an example embodiment, risk management system 6102 may be configured to manage risk or liability for good or good components.

In an embodiment, machine learning system 6150 may be used by artificial intelligence system 1160 to classify, predict, and/or make other decisions related to risk management, including for product 650 and product components. Train one or more models (6120). In an embodiment, they may be equipment components. In an example embodiment, model 6120 is trained to mitigate risk and liability by detecting the state of a set of components. Machine learning system 6150 can train a model using an n-tuple containing one or more results associated with features about a component and component conditions. In this example, the characteristics for a component include component material (plastic, glass, metal, etc.), component history (manufactured date, usage history, repair history), component characteristics, component dimensions, component thermal properties, component price, component supplier, and /or other related features may be included, but are not limited thereto. In this example, the results may include whether the digital twin of component 6002 is in operating condition. In this example, one or more characteristics of the digital twin may be varied for different simulations, and the results of each simulation may be recorded in a tuple with relevance. Other examples of trained risk management models may include models 6120 that are trained to optimize product safety, models that are trained to identify components that are likely to cause undesirable events, etc.

In operation, artificial intelligence system 1160 may use the model 6120 discussed above to make risk management decisions on behalf of risk management system 6102, given one or more characteristics related to a task or event. For example, artificial intelligence system 1160 can determine the condition of a component. In this example, risk management system 6102 may provide characteristics of the component to artificial intelligence system 1160. These characteristics may include component material, component history, component dimensions, component cost, component thermal characteristics, component supplier, etc. In embodiments, artificial intelligence system 1160 may feed these features to one or more of the models discussed above to obtain one or more decisions. This determination may include whether the component is in operating condition.

In embodiments, artificial intelligence system 1160 may utilize component digital twin 6002 to run simulations for proposed decisions.

Risk management system 6102 may then use the decision(s) made by artificial intelligence system 1160 to initiate a component resupply event. Additionally, following a component resupply event, the results of the event (e.g., improved product performance) may be reported to machine learning system 6150 to strengthen the model used to make decisions.

Platform 604 may include, integrate, be integrated with, manage, control, coordinate, or otherwise handle component attribute digital twin 6140. Component attribute digital twin 6140 may be a digital representation of a set of attributes of a set of supply chain components in the supply of an enterprise's set of products.

An environment for the development of a component attribute digital twin 6140 allows developers to take input from selected data sources in the data storage layer 624 and events or other data from the monitoring system layer 614 and compile them into a component attribute digital twin 6140. ) may include a set of interfaces for developers to configure the artificial intelligence system 1160 to provide for inclusion in Component Properties The digital twin development environment can be configured to take output and results from various applications 630. In embodiments, component attribute digital twin 6140 may be provided for the broad value chain network application 630 mentioned throughout this disclosure and documents incorporated by reference herein.

In embodiments, methods, systems, and devices include a value chain network management platform having an asset management application associated with maritime assets, and a set of maritime activities of one or more of the maritime assets and design results, parameters, and parameters associated with the one or more maritime assets. An information technology system having a data handling layer of a management platform that includes data sources containing information used to populate a training set based on one of the data. The information technology system is also configured to learn based on the training set collected from the data source, simulate one or more properties of one or more of the marine assets, and change the one or more properties based on the training set collected from the data source. It has an artificial intelligence system that generates one or more sets of recommendations. The information technology system may also be included in a value chain network management platform that provides visualization of a digital twin of one or more marine assets, including details generated by an artificial intelligence system of one or more attributes in combination with one or more recommendation sets. It has a digital twin system.

6, the value chain network management platform 604 provides information on the value chain network 668 when applied to maritime assets, activities, logistics, and planning, including supply and production factors, demand factors, logistics and distribution factors, etc. Orchestrates the various factors involved in planning, monitoring, controlling, and optimizing the various entities and activities involved. Management platform 604 monitors and manages supply factors and demand factors, when demand factors are understood and accounted for, when orders are created and processed, and when products are created and moved through the supply chain. And sharing of status information between them can be facilitated. Referring to FIG. 7 , management platform 604 may include a set of value chain network entities 652 that include various delivery systems 632 that may include and connect to marine facilities 622 . Marine facilities 622 may include port infrastructure facilities 660, floating assets 620, and shipyards 638, etc. In embodiments, the value chain network management platform 604 provides management (and in some cases autonomous or semi-autonomous behavior) of a wide range of value chain network 668 processes, workflows, activities, events and applications 630 applicable in a maritime environment. ) to monitor, control, and otherwise enable.

6 and 11, a management platform 604 deployed in a marine environment may include a set of data handling layers 608, each of which may be used for a wide variety of value chain network applications and It is structured to provide a set of capabilities that facilitate the development and deployment of intelligence, to facilitate the application of automation, machine learning, artificial intelligence, intelligent transactions, state management, event management, process management, etc. for end use. . In an embodiment, the data handling layer 608 is structured in a topology that facilitates the collection and distribution of shared data across multiple applications and uses within the management platform 604 by the value chain monitoring system layer 614. Value chain monitoring system layer 614 is for collecting and organizing data collected from or about value chain entities 652, as well as data collected from or about various data layers 624 or services or components. It may include, integrate with, and/or cooperate with various data collection and management systems 640, referred to in some cases for convenience as data collection system 640.

In an embodiment, the data handling layer 608 is connected to the platform 604 by a value chain network-oriented data storage system layer 624, which in some cases is referred to herein for convenience simply as the data storage layer 624 or storage layer 624. ) consists of a topology that facilitates sharing or common data storage across multiple applications and uses. For example, various data collected about value chain entities 652 as well as data generated by other data handling layers 608 may be stored in data storage layer 624 so that various data handling layers 608 Any of the services, applications, programs, etc. may access a common data source, which may include a single logical data source distributed across disparate physical and/or virtual storage locations. This facilitates a dramatic reduction in the amount of data storage required to handle the enormous amounts of data generated by or about value chain network entities 652 as applications 630 and use of the value chain network grow and proliferate. can do. For example, a supply chain or inventory management application in the value chain management platform 604, such as for ordering replacement parts for an item of machinery or equipment, may determine the likelihood that a ship's component or a port's facility will require a replacement part. Predictive maintenance applications can access the same data set about which parts have been replaced for a set of machines. Similarly, forecasts may be used in connection with resupply of items.

6 and 12, the value chain network-oriented data storage system layer 624 includes, without limitation, a physical storage system, a virtual storage system, a local storage system 1190, a distributed storage system, a database, memory, and a network. May include storage-based and network-attached storage systems. In an embodiment, the storage layer 624 may be associated with one or more knowledge graphs within the graph database architecture 1124, such as directed acyclic graphs, data maps, data hierarchies, data clusters containing links and nodes, self-organizing maps, etc. Data can be saved. In embodiments, the data storage layer 624 may include any of the entities described in this application, e.g., to maintain serial or other records of entities 652 over time, such as digital threads, ledgers, distributed Data can be stored in ledgers, etc. In an embodiment, storage layer 624 stores identity data, for example, with access controls that may be role-based or based on credentials associated with value chain entity 652, service, or one or more applications 630. , may include one or more blockchains 1180, such as storing transaction data, past interaction data, etc. Data stored by data storage system 624 may include accounting and other financial data 730, access data 734, asset and facilities data 1032, asset tag data 1178, operator data 1032, event data ( 1034), risk management data 732, price data 738, safety data 664, etc.

Referring to FIG. 8 , the value chain network management platform 604 includes one or more sets of value chain entities 652, the value chain entities to which management by the management platform 604 may be applied, and the management platform 604 ), and/or may supply inputs to and/or take outputs from the management platform 604, such as for or in conjunction with broader value chain activities. These value chain entities 652 may include any of a wide variety of assets, systems, devices, machines, components, equipment, facilities, and individuals that can support a wide range of operational facilities 712, including offshore facilities 622. You can. Referring to FIG. 63, marine facilities may include port infrastructure facilities 7000. In embodiments, port infrastructure facility 7000 includes docks 7002, yards 7004, cranes 7008, roll-on/roll-off facilities 7010, ramps 7012, and containers, as applicable. 7014, a container handling system 7018, a waterway 732, and a locking device 7020. In an embodiment, dock 7002 and adjacent areas may include dock 7022, water basin 7024, stacking area 7028, storage area 7030, and warehouse 7032. In embodiments, container handling system 7018 may include one or more containers for moving shipping containers, such as cranes (e.g., Gottwald cranes, gantry cranes, etc.), straddle carriers, multitrailers, reach stackers, etc. Can include container tracking systems and sensors 7040 for monitoring, reporting, or managing other systems. In embodiments, port infrastructure facility 7000 may further include gantry cranes 7042 and port vehicles 7044 that may be used to move containers 7014, such as straddle carriers. In an embodiment, the port infrastructure facility 7000 also includes a refrigerated container 7050 with a dedicated stacking area 7052 and cooling infrastructure to maintain a controlled environment within the refrigerated container 7050.

Port infrastructure facilities (7000) further include shipyard facilities (638) and floating assets (620). Floating assets (620) include ships (7060) and boats, container ships (7062), barges (7064), tugboats (7068, 7070), and small boats (7072), as well as partially floating assets such as submarines, underwater drones, etc. can do. In this example, floating assets 620 may operate between facilities and other items at an origin point 610 and/or a destination point 628. Shipyard facility 638 may include a transportation facility 710 such as a number of floating assets 620 as well as land-based vehicles and other delivery systems 632 used to transport goods such as trucks, trains, etc.

63, the orchestration of a set of deeply interconnected value chain network entities 652 by the management platform 604 includes local network connections, peer-to-peer connections, connections over one or more mobile networks, and cloud networks. This may include providing interconnectivity to value chain network entities 652 using facility-based connections, satellite uplinks, microwave communications, or other connections. Management platform 604 manages connectivity and provides connectivity, recognizing that floating assets 620 are in many maritime environments where connectivity may be poor or non-existent compared to when closer to a port or other land-based communications system. You can configure or provision resources to enable, and/or manage the application 630 using connections. In many instances, a port infrastructure facility 660, such as a yard for holding shipping containers 7080, may notify the fleet of floating assets 620 via a connection to a floating asset 620 that the port is approaching capacity limits. You can. With this knowledge, floating asset 620 movement can be modified to extend the time, including reducing approach speed to delay arrival, direction to another port, etc. In another example, news of a port capacity limit being reached may cause a negotiation process to begin with floating assets 620 to reach the port. In an embodiment, the negotiation process with floating assets 620 is based on a set of rules and governed by smart contracts for remaining capacity and is automated to allow some floating assets 620 to be redirected to alternative ports or holding facilities. May include negotiated negotiations.

In embodiments, marine facility 622 may include floating assets 620 that include many different vessels 7060. 64 and 65, vessel 7060 may be one or more container ships 7062 capable of carrying many shipping containers 7080. In another example, vessel 7060 may be one or more container ships 7062 capable of carrying raw materials, bulk processed goods, gas cargo, and many other types of cargo not otherwise transported within shipping containers 7080. In many examples, vessel 7060 may include bow area 7100. Bow area 7100 may include a bulbous bow 7102. In some examples, concept player 7102 may be configured on-site in response to controls from management platform 604. Moving internally from the bow area 7100 toward the stern area 7104 of the vessel 7060, the vessel 7060 may include a forepeak tank 7110. In this same area, watercraft 7060 may include one or more bow anchors 7112 and bow thrusters 7114. Various passageways 7118 connect these areas to bow area 7100. Depending on the configuration of vessel 7060, holds 7120 may be configured and reconfigured to accommodate a variety of products, such as product 1510, raw materials, in-process materials, and combinations thereof. In some examples, vessel 7060 may include multiple holds 7120. In an example, container ship 7062 may be comprised of eight holds: container holds 7130, 7132, 7134, 7138, 7140, 7142, 7144, 7148. Toward the aft area 7104, vessel 7060 includes an engine room 7150 containing one or more propulsion units 7152. Each of one or more propulsion units 7152 is supplied by a fuel system 7154 and its emissions are controlled by an exhaust system 7158. At various locations on the vessel 7060, one or more fin stabilizers 7160 may be placed. In the aft area 7104, the watercraft 7060 includes a steering gear area 7160 below the aft deck area 7162. One or more rudders 7164 may extend from steering gear area 7160.

One or more propellers 7170 may extend from the aft area 7104 to the propulsion unit with a rotating power connection. In embodiments, one or more propellers 7170 may extend from vessel 7060 with electrical connections to propulsion units but no physical rotational power connections. In embodiments, one or more propellers 7170 may extend from vessel 7060 to the propulsion unit with a hydraulic connection but no physical rotational power connection. In another example, steam or other working fluid may be used to drive propulsion of vessel 7060. In other examples, mechanical rotary power, electric drives, hydraulic drives, steam, and various combinations thereof may be used for propulsion. In various examples, one or more propellers 7170 may include side propellers 7172 and center propellers 7174. In another example, two propellers 7170 may be deployed. In embodiments, propeller 7170 may be fixed such that the plane in which the propeller rotates is fixed relative to vessel 7060. In this example, propeller 7170 may be fixed and driven by a mechanical link to a propulsion unit of vessel 7060. In another example, the propellers 7170 may be fixed and driven by an electric motor adjacent to each of the propellers 7170. In embodiments, the position of the propeller 7170 may be variable such that the plane in which the propeller rotates is movable relative to the vessel 7060. In this example, propellers 7170 may be driven by electric motors adjacent to each propeller 7170. At one or more locations on vessel 7060, propellers 7170 may be placed within a pod that may include independently controlled and movable electric drivetrains and propellers, such that the entire pod may be capable of forward propulsion, steering, maneuvering, and docking. , it can be moved to various locations to facilitate evasive maneuvers, etc.

In another example, vessel 7060 is comprised of one or more ballast tanks 7180. In various examples, vessel 7060 may include side ballast tanks 7182 and deep ballast tanks 7184. Ballast tanks 7180 are configured to control ballast water characteristics such as pumping and drainage system 7190, cleaning system 7192, salinity, debris, organic matter, garbage, restricted contents for geofenced areas, regulated areas, ad hoc boundary areas, etc., respectively. It may include a sensor 7194 to determine . Sensor 7194 may also determine tank characteristics including wear from fatigue, corrosion, physical damage, etc. In the bow area 7100, the vessel 7060 has a windlass where various sensors 7208 can be located to observe the characteristics of the vessel 7060, weather and ambient conditions 7210, and navigation inputs 7212. windlass) (7200), foremast (7202), and crow-nest (7204). At various locations on the vessel 7060, one or more mooring winches 7220 may be positioned to assist docking with respect to a suitable mooring connection point, connection to another moving vessel such as a tender, etc. At various locations on vessel 7060, one or more hatch covers 7222 may be positioned to allow access to various areas and passageways on vessel 7060.

In another example, vessel 7060 is configured as a container vessel 7062, which may be comprised of eight holds: container holds 7130, 7132, 7134, 7138, 7140, 7142, 7144, 7148. In another example, vessel 7060 is configured as a container vessel 7062 with various numbers of holds 7120. In another example, vessel 7060 is configured as a container vessel 7062 with field-configurable holds. In another example, vessel 7060 is configured as a container vessel 7062, a portion of which has a variable number of field-configurable holds. In embodiments, holds 7120 may include one or more vents 7240 deployed to facilitate an atmosphere within the hold suitable for movement and management of cargo. In embodiments, hold 7120 may include one or more rigs and anchoring systems 7242 to secure one or more loads within hold 7120 configured or reconfigured for such loads. In embodiments, holds 7120 may include one or more movable baffles and bedding 7244 to secure one or more loads within holds 7120 that are configured or reconfigured for such loads.

In another example, watercraft 7060 includes a wheelhouse 7250 and one or more lifeboats 7252 and 7254. In another example, vessel 7060 includes marine and satellite navigation equipment 7260. In this example, the vessel may include a direction finder antenna 7262, a radar scanner 7264, and a signal yard 7268. In this example, vessel 7060 includes a radar mast 7270 and a Suez signal light 7272, a funnel 7274, and an antenna pole 7278.

In another example, vessel 7060 includes one or more cranes 7280 that can be used to move objects within and around deck 7282 of vessel 7060 and in and out of holds 7120. In this example, vessel 7060 can accommodate or carry on top many containers of various sizes, including 20-foot and 40-foot containers. In this example, vessel 7060 can accommodate or carry on top a number of containers of various sizes, including a 20-foot dry cargo container, a 20-foot open top container, a 20-foot collapsible flat rack container, a 20-foot refrigerated container, etc. You can. In this example, vessel 7060 may accommodate on top a number of containers of various sizes, including 40-foot tall cube containers, 40-foot open top containers, 40-foot collapsible flat rack containers, 40-foot tall cube refrigerated containers, etc. It can be transported. In this example, vessel 7060 can accommodate or carry on top many containers of various sizes, including 45-foot tall cube dry containers, etc.

In embodiments, vessel 7060 may include an engine unit that includes a diesel generator 7280 that may provide power throughout vessel 7060. Vessel 7060 may also include engine units including a central main diesel engine 7282 and one or more side main diesel engines 7284. In embodiments, watercraft 7060 may include an engine unit configured to burn natural gas, propane, gasoline, methanol, etc. In embodiments, watercraft 7060 may include an engine unit configured to be driven by a nuclear unit that may be used to heat water to a steam driven electrical system. In embodiments, watercraft 7060 may include a nuclear unit and an engine unit configured to be driven by an internal combustion engine in a hybrid arrangement. In embodiments, the vessel 7060 may include a nuclear unit and an engine unit configured to be driven by an internal combustion engine, and other renewable energy sources in a hybrid arrangement, such as solar and wind, where each of these can provide power propulsion and the vessel. It can supply electricity and battery systems for operation.

In embodiments, vessel 7060 may include multiple bulkheads 7290. In this example, the engine room may be framed within engine room bulkhead 7292 to contain various powerplant units. In embodiments, the cargo and hold areas of vessel 7060 may include a hold bulkhead 7294 to contain various power plant units. In embodiments, watercraft 7060 may include structural transverse bulkheads 7300 and axial bulkheads 7302.

In embodiments, marine facility 622 may include floating assets 620 that include many different barges 7500 . Referring to FIG. 66, one or more barges 7500 may be transport barges, cargo barges, submersible barges, etc., of any size and capacity. In many instances, barges are available in a variety of self-propelled vessels, including many different tow barges and submersible heavy lift vessels. In many examples, barge 7500 may be towed or propelled by tow boat 7510 for transport from one location to another. In many examples, barge 7500 may have a flat top and bottom and may be equipped with navigation lights 7520, fairleads 7522, and towing points 7524.

In some examples, barge 7500 may be designed to be submerged to pick up cargo 7530, such as floating cargo. In this example, barge 7500 may have a cabin 7540 and a deck structure 7542 in the bow area 7550 opposite a deck structure 7544 in the aft area 7552. There may be an additional deck structure 7548 between the bow area 7550 and the aft area 7552 that can be configured and reconfigured to hold cargo 7530. In this example, barge 7500 may be equipped with its own ballast system 7560. In an embodiment, barge 7500 has modular steel boxes 7570 and deck structures 7542, 7544, 7548 that enhance the stability of barge 7500 and its cargo 7530 when passing through waterline 7584. A stability casing 7572 may be added to the stern area 7552 to some predetermined extent to effectively provide additional portion of the hull 7580 within the water 7582 that may appear to be present. In this example, modular steel box 7570 and stability casing 7572 may be removable and stowed on one of the deck structures 7542, 7544, 7548 of barge 7500 or stored onshore when not required. You can. By doing so, barge 7500 can be relatively more efficient as lighter loads warrant a relatively smaller hull structure.

In many instances, barges 7500 may be classified by their length and width, as well as how they are used, launched, etc. In some examples, one or more barges 7500 may be less than 200 feet long and 50 feet wide. In this example, barge 7500 may include a small pontoon that can be used to transport small structures in protected coastal waters. In some examples, one or more of the barges 7500 may be approximately 250 feet by 70 feet and may include small pontoons to support barges 7500 that are otherwise configured without an onboard ballast system. In this example, barges in this configuration may be used to transport small marine loads, work in and near port infrastructure, perform maintenance at shipyards, etc. In some examples, one or more of the barges 7500 may be approximately 300 feet long and 90 or 100 feet wide. In this example, one or more barges of this configuration may be utilized as standard cargo barges, but may not be equipped with an onboard ballast system. In some examples, one or more barges 7500 may be approximately 400 feet by 100 feet, and these barges may be equipped with an onboard ballast system.

In some examples, one or more barges 7500 may be approximately 450 feet or longer and may be deployed with an onboard ballast system 7590. In this example, one or more barges 7500 may also be deployed with skid beams 7592. One or more barges 7500 may also be disposed with rocker arms 7594 in the aft area 7552, for example, to enable launch of jackets or other loads that may be too heavy to lift. In the example, the Heerema H851 brand barge is nominally 850 feet long by 200 feet wide and may be a suitable example of one of the largest commercially available barges.

In embodiments, one or more barges 7500 may be configured as submersible barges 7600, which may also be towing barges that may be equipped with a stability casing 7602 in the stern area 7552. In an example, submersible barge 7600 may be configured with a ship-like bow structure 7604. In this example, a vessel such as bow structure 7604 may be configured with a bridge 7608 high enough to enable submersion of the barge over at least a portion of its deck structure. In the example, a BOA brand barge has nominal dimensions of 400 feet by 100 feet, an AMT brand barge has nominal dimensions of 470 feet by 120 feet, and a Hyundai brand barge with nominal dimensions of 460 feet by 120 feet is commercially available. These may be suitable examples of possible submersible barges. In this example, this barge is capable of submerging 18 to 24 feet above its deck.

In light of the present disclosure, it will be understood that barges are rated and paired with jobs in terms of dead weight, which provides a broad indication of the barge's carrying capacity. However, barges have additional requirements such as global strength, local deck and frame strength and the height of the cargo's center of gravity. With regard to center of gravity, one exemplary barge would be capable of transporting a 20,000 ton structure, the center of gravity of which would be very close to the deck so that it is sufficiently anchored and supported on the deck. The same exemplary barge can transport half the weight only if the cargo has a relatively high center of gravity. With this in mind, many attributes of one or more barges are the placement, orientation, center of gravity, and weight of cargo on its decks.

In an embodiment, one of the barges is a towing bridle 7610 that can be towed by one of a ship, a tugboat 7510, etc. In many examples, two lines 7612 may extend from the towing bracket 7614 through the fairlead 7618 on one of the barges and connect to the triplate 7620 on the barge via the towing shackle 7622. In this example, third line 7630 may connect triplate 7620 to winch 7640 on one of the tugboats 7510. In another example, emergency wires 7642 may be installed along the length of the barge. Emergency wire 7642 may be attached to connector 7644, which may terminate in buoy 7650. Buoy 7650 may follow barge 7650 during towing and may form part of a towing array.

In some examples, the roll acceleration of a barge may be directly proportional to the lateral stiffness of the barge, which can be measured by its transverse height. In some arrangements, the barge may have a large radial height and, as a result, roll acceleration may be significant. In another example with a relatively high load, the radial height may be low, resulting in a period and amplitude of roll and a larger static force resulting from the load, but a smaller dynamic component. In many examples, the attributes of barge 7500 include the positioning of cargo 7530 on its deck structure and its effective midship height. In another example, counter-roll mechanism 7660 may be installed on barge 7500. In this example, adaptive intelligence layer 614 may update the program of counter-roll mechanism 7660 and appear to increase efficacy against changing cargo loads and water and weather conditions. In embodiments, adaptive intelligence layer 614 may update the speed and angle of counter-roll mechanism 7660, resulting in increased efficacy for changing cargo loads and water and weather conditions.

In embodiments, management platform 604 may include a set of value chain network entities 652 that include various delivery systems 632 that may include and connect to marine facilities 622 . Marine facilities 622 may include port infrastructure facilities 660, floating assets 620, and shipyards 638, etc. In embodiments, the value chain network management platform 604 provides management (and in some cases autonomous or semi-autonomous behavior) of a wide range of value chain network 668 processes, workflows, activities, events and applications 630 applicable in a maritime environment. ) to monitor, control, and otherwise enable.

Marine facility 622 may include one or more vessels 7060 of various sizes to service the facility. Marine facility 622 may include one or more navigation aids anchored or moored in the water or on land to facilitate movement of vessels and land vehicles of various sizes. In embodiments, marine facility 622 may be configured as a port in that it may be configured to accommodate deep draft vessels with drafts of 20 feet or more. In embodiments, some of the larger marine facilities 622 may include areas outside the boundaries of ports, shipyards, marine ports, etc., associated with port operations or intermodal connections to ports, shipyards, marine ports, etc.

In embodiments, management platform 604 may manage port gate-in and gate-out improvements to the logistics of the flow of assets and cargo around marine facility 622. In embodiments, management platform 604 may manage roadway improvements both within and connected to marine facility 622. In embodiments, management platform 604 may manage rail improvements both within and connected to marine facility 622. In embodiments, management platform 604 may manage marina improvements at marine facility 622, including docks, wharves, docks, etc. In embodiments, the management platform 604 may manage marina improvements, including dredging at the marina, approach and departure areas adjacent to the marina, and areas around marine facilities. In embodiments, management platform 604 may manage cargo movement equipment used on land. In embodiments, the management platform 604 may be used to manage silos, elevators, conveyors, container terminals, roll-on/roll-off facilities including parking for intermodal cargo transportation, warehouses including refrigeration facilities, and storage facilities for oil or gas products. We can manage facilities needed to improve freight transport, including bunkering facilities, lay-down areas, transit sheds, etc. In embodiments, management platform 604 may manage utilities necessary for standard operation, including lighting, stormwater, etc., that may be incidental to a larger set of marine facilities. In embodiments, management platform 604 provides efficient port management that includes flow-through processing for import/export requirements, storage and tracking, and asset/equipment management, as well as routing and communications for ship, truck, and rail freight movements. Manage port-related intelligent transport system hardware and software, including all technologies used to facilitate movement. In embodiments, management platform 604 may manage phytosanitary treatment facilities to support phytosanitary treatment requirements. In embodiments, management platform 604 may manage, configure, and reconfigure fully automated cargo handling equipment.

In embodiments, the adaptive intelligence system layer 614 is a system, components, and systems that collectively facilitate the coordinated development and deployment of intelligent systems, such as those capable of enhancing one or more of the applications 630 on the application platform 604. , a set of services and other capabilities; capable of improving the performance of one or more components, or the overall performance of the connectivity facility 642 (e.g., speed/latency, reliability, quality of service, cost reduction, or other factors); capable of improving other capabilities within the adaptive intelligence system layer 614; improving the performance (e.g., speed/latency, energy utilization, storage capacity, storage efficiency, reliability, security, etc.) or overall performance of one or more components of the value chain network-oriented data storage system 624; controlling, automating, or optimizing one or more performance characteristics of one or more value chain network entities 652; or generally improving any of the processes and application outputs and results 1040 sought by use of the platform 604.

These adaptive intelligence systems 614 may be deployed within and between marine facilities 622 and floating assets 620 . This adaptive intelligence system 614 includes a robotic process automation system 1442, a set of protocol adapters 1110, a packet acceleration system 1410, an edge intelligence system 1430 (which may be a self-adaptive system), an adaptive A networking system 1430, a set of state and event managers 1450, a set of opportunity miners 1460, a set of artificial intelligence systems 1160, a set of digital twin systems 1700, (e.g., a value chain network 668 ), and other systems for setting up, provisioning, configuring, and otherwise managing the set of interactions between the set of value chain network entities 652 within the network. .

In an embodiment, a set of digital twin systems 1700 may be deployed for each of the marine facilities 622 and each of the floating assets 620. Referring to FIG. 6, the connected value chain network 668 plans, monitors, controls, and controls various entities and activities accompanying the value chain network 668, such as supply and production factors, demand factors, logistics and distribution factors, etc. and a digital twin system deployed throughout the value chain network management platform 604 to facilitate the management, visualization, and modeling of the orchestration of the various factors involved in optimizing. By means of an integrated platform 604 for monitoring and managing supply factors and demand factors, a digital twin of status information facilitates modeling and analysis, as orders are created and processed, and as products are created and supply chain As they move through, they can be shared across and across various entities to provide visualization of changing demand factors that become operational realities.

In embodiments, value chain monitoring system layer 614 and data collection system 640 may include a wide range of systems for collection of data from marine facilities 622 and floating assets 620. This layer may include, without limitation, real-time monitoring systems 1520 (e.g., vessels and other floating assets, delivery vehicles, trucks and other transport assets, and event and status reporting systems at shipyards, ports, warehouses, distribution centers and other locations). such as on-board monitoring system); On-board diagnostics (OBD) and telematics systems for floating assets, vehicles and equipment; A system that provides diagnostic codes and events via an event bus, communications port, or other communications system; as well as monitoring infrastructure (such as cameras, motion sensors, beacons, RFID systems, smart lighting systems, satellite connectivity, asset tracking systems, people tracking systems, and ambient sensing systems located in various environments where value chain activities and other events occur). as well as removable devices for maritime assets and cargo or other assets contained within or moving on them, such as portable and mobile data collectors, RFID and other tag readers, smartphones, tablets and other mobile devices capable of data collection. and replaceable monitoring systems; a software interactive observation system 1500 that can be deployed on portable and onboard systems on marine facilities 622 and floating assets 620; To detail cargo in the hold of floating asset 620, to detail the activities of personnel and gear deployed at marine facility 622 and on floating asset 620, such as items, people, materials, components, machinery , video and still imaging systems that allow visualization of equipment, personnel, etc., visual monitoring systems using LIDAR, IR, and other systems (1930); Interaction point systems (e.g., dashboards, user interfaces, and control systems for value chain entities); Physical process observation system 1510 (e.g., physical activities of operators, workers, customers, etc., individuals (e.g., shippers, delivery workers, packers, pickers, assembly personnel, customers, merchants, vendors, distributors, and others)) For tracking physical activity, the physical interaction of workers with other workers, the interaction of workers with physical entities such as machines and equipment, and the interaction of physical entities with other physical entities, including, without limitation, video and still image cameras; Including through the use of motion detection systems (e.g., including optical sensors, LIDAR, IR and other sensor sets), robotic motion tracking systems (e.g., tracking the movement of systems attached to humans or physical entities), etc. ; Machine health monitoring system 1940 (any value chain entity, e.g., machine or component, e.g., machine, e.g., client, server, cloud resource, control system, display screen, sensor, camera, vehicle, robot, or other machine) (including on-board monitors and external monitors of conditions, states, operating parameters, or other measures of conditions); Sensors and Cameras 1950 and other IoT data collection systems 1172, including onboard sensors, value chain environments (e.g., without limitation, points of origin, loading or unloading docks, vehicles or floating assets used to transport goods, containers, Sensors or other data collectors (including click tracking sensors) in or around ports, distribution centers, storage facilities, warehouses, delivery vehicles, and destination points), cameras to monitor the overall environment, and certain Dedicated cameras for machines, processes, workers, etc., wearable cameras, portable cameras, cameras placed on mobile robots, cameras on portable devices such as smartphones and tablets, and cameras incorporated by reference throughout this disclosure or in this application. including many others including any of the many sensor types disclosed in the document); Indoor Location Monitoring System 1532 (cameras, IR systems, motion detection systems, beacons, RFID readers, smart lighting systems, triangulation systems, RF and other spectral detection systems, time-of-flight systems, chemical noses, and other chemical sensor sets) but also includes other sensors); User feedback system 1534 (including survey system, touch pad, voice-based feedback system, evaluation system, facial expression monitoring system, affect monitoring system, gesture monitoring system, etc.); Behavioral monitoring system 1538 (e.g., movement, shopping behavior, purchasing behavior, clicking behavior, behavior indicative of fraud or deception, user interface interaction, product return behavior, behavior indicative of interest, attention, boredom, etc., mood indications) to monitor behavior (e.g., fidgeting, staying still, moving closer, changing posture, etc.); and a wide variety of Internet of Things (IoT) data collectors 1172 such as those described throughout this disclosure and in documents incorporated by reference herein.

26, a set of opportunistic miners 1460 may be provided as part of an adaptive intelligence layer 614, which may include, for example, the addition of artificial intelligence 1160, automation (robotic process automation 1402), ), etc., to one or more of the marine facilities 622 and to each of the floating assets 620, including the systems, subsystems, components, and applications with which the platform 100 interacts, among the elements of the platform 604. It can be configured to find and recommend opportunities to improve one or more things. In embodiments, opportunity miner 1460 may be configured or used by developers of AI or RPA solutions to find opportunities for better solutions and optimize existing solutions in value chain network 668. In an embodiment, Opportunity Miner 1460 is a system that collects information within Management Platform 604 and information within, about, and about a set of Marine Facilities 622 and for each of Floating Assets 620. may include a set of, wherein the collected information relates to increased automation and/or one or more of the value chain network 668, the application 630, the offshore facility 622, and the floating asset 620. Or, it has the potential to help identify and prioritize opportunities for intelligence. For example, opportunity miner 1460 may identify labor-intensive areas and processes in a set of value chain network 668 environments, e.g., by time, type, using cameras, wearables, or other sensors. It could include a system that observes clusters of value chain network workers by location, whether on water or on land. These can be used to show places with high labor activity, for example in a ranked or prioritized list, or in visualizations (for example, heat maps showing the dwell time of customers, workers or other individuals on a map of the environment or customers within the environment). Or it can be presented in a heat map showing the path the worker took. In embodiments, analytics 838 may be used to identify which environments or activities would most benefit from automation for the purposes of improved delivery times, alleviation of congestion, and other performance improvements.

In embodiments, opportunity mining may include facilities for requesting appropriate training data sets that may be used to facilitate process automation. Certain types of input, if available, will provide very high value for automation, for example, video data sets that capture very experienced and/or highly specialized workers performing complex tasks. This information becomes even more valuable when collected in close proximity to other offshore facilities 622 and in conjunction with deployed floating assets 620. Opportunity miner 1460 may search such video data sets as described herein; However, in the absence of success (or to supplement available data), management platform 604 may provide the types of data desired by users at a maritime facility or deployed on a maritime asset, such as software interaction data (e.g., to perform specific tasks). of experts working with a program to perform them), video data (e.g. a video showing a set of experts performing a particular kind of delivery process, unloading process, security and logistics process, cleaning and maintenance process, container movement process, etc. ), and/or a system capable of specifying physical process observation data (e.g., video, sensor data, etc.). The resulting library of interactions captured in response to the specifications is stored as a data set in the data storage layer 624, for example, for consumption by various applications 630, adaptive intelligence system 614, and other processes and systems. can be captured. In embodiments, the library provides automation maps that can follow instructions, e.g., in a video, such as providing a sequence of steps according to a procedure or protocol, decomposing a procedure or protocol into sub-steps that are candidates for automation, etc. In order to facilitate the development of a video, it may include a video that is specifically developed as an educational video. In embodiments, such videos may be used to automatically develop sequences of labeled instructions that developers can use, for example, to facilitate maps, graphs, or other models of the process to aid in the development of automation for the process; It can be processed by natural language processing.

In an embodiment, the value chain monitoring system layer 614 and data collection system 640 include an entity discovery system for discovering one or more value chain network entities 652, such as any of the entities described throughout this disclosure. 1900), and these entities may include, among other things, those that may be loaded and offloaded as control passes from various marine facilities 622 and floating assets 620. This can be done by device identifier, by network location, by geolocation (e.g. by geofencing), by indoor location (e.g. by proximity to known resources such as IoT-enabled devices and infrastructure, Wifi routers, switches, etc.). ), by cellular location (e.g., proximity to a cellular tower), by marine navigation aids and vessel identity beacons, by an identity management system (e.g., an entity 652 is assigned by platform 604 Entities in offshore facilities 622 and floating assets 620 within the value chain network 668, such as when associated with another entity 652, such as an owner, operator, user, or corporation, and/or by a managed identifier. It may include components or subsystems for searching. In this example, entity discovery system 1900 may interact with an established maritime asset logistics system used to track traffic and location. In this example, entity discovery system 1900 may interact with established maritime asset autopilot and automatic navigation systems to obtain information related to the intended navigation destination and, therefrom, determine error correction actions necessary to reach the navigation destination. and obtain the size.

22, the adaptive intelligence layer 614 is a set of components, processes, services, interfaces and other elements for the development and deployment of digital twin capabilities for visualization of various value chain entities 652 in the environment. , and applications 630, as well as other added value enabled or facilitated by coordinated intelligence and digital twins 1700 (including artificial intelligence 1160, edge intelligence 1420, analytics and other capabilities). May include a value chain network digital twin system 1700 that may include services and capabilities. In embodiments, digital twin system 1700 may be deployed with each facility (or group thereof) of marine facilities 622 and may be deployed for each of floating assets 620 . In many cases, each floating asset 620 and physical asset within marine facility 622 can be coordinated and managed with its digital twin supported by digital twin system 1700. Without limitation, digital twin system 1700 may be a platform application layer that can be deployed across (or across groups of) various systems, networks, and infrastructures of floating assets 620 and within and between marine facilities 622 A set of applications 614 may be used and/or applied to each of the processes managed, controlled, or mediated by each.

In embodiments, digital twin 1700 may take advantage of the presence of multiple applications 630 within value chain management platform 604, such that a pair of applications can perform fusion of collected signals, etc., for value chain entities 652. (e.g., from the data storage layer 624) and other inputs (e.g., from the monitoring layer 614), as well as shared data sources (e.g., from the data storage layer 624), as well as throughout this disclosure and references therein. and monitoring through the use of artificial intelligence 1160, including any of the various expert systems, artificial intelligence systems, neural networks, supervised learning systems, machine learning systems, deep learning systems, and other systems described in the incorporated documents. Outputs, events, state information and outputs may be shared, including through the use of content collected by layer 614 and data collection system 640, which may be used to enrich content in digital twin 1700. It can collectively provide a much richer environment.

23 , as described throughout this disclosure, a component or element of the value chain network entity 652, application 630, or platform 604, e.g., event data 1034, status, etc. By populating digital twin 1700 with value chain network data objects 1004, such as data 1140 or other data, any of the value chain network entities 652 are represented in a set of one or more digital twins 1700. It can be.

Accordingly, platform 604 may include, integrate, be integrated with, manage, control, coordinate, or otherwise handle any of the following, among others: distribution twin 1714 (e.g., A wide variety of digital twins (1700), such as those representing distribution facilities, assets, objects, workers, etc.; Warehousing Twin (1712) (e.g., representing warehouse facilities, assets, goods, workers, etc.); port infrastructure twin 1714 (e.g., representing a port, airport, or other facility, as well as assets, objects, workers, etc.); Shipping Facility Twin (1720); Operating Facility Twin (1172); Customer Twin (1730); Worker Twin (1740); Wearable/Portable Device Twin (1750); Process Twin (1760); machine twin 21010 (e.g., for the various machines used to support the value chain network 668); product twin (1780); point of origin twin (1502); provider twin (1630); supplyfactor twin(1650); Marine Facility Twin (1572); Floating Asset Twin (1570); Shipyard Twin (1620); Destination Twin (1562); Fulfillment Twin (1600); Delivery System Twin (1610); demand factor twin (1640); Retailer Twin (1790); E-commerce and online site and operator Twin (1800); Aqueduct Twin (1810); Road Twin (1820); Railway Twin (1830); aviation facility twins 1840 (e.g., twins of aircraft, runways, airports, hangars, warehouses, air travel corridors, refueling facilities, and other assets, objects, personnel, etc. used in connection with the air transportation of products 650); Autonomous Vehicle Twin (1850); Robotics Twin (1860); Drone Twin (1870); and Logistic Factor Twin (1880); etc.

Referring to FIG. 27 , platform 604 relates to elements of an adaptive intelligence layer 614 that facilitates improved edge intelligence, including, in particular, an adaptive edge computing management system 1400 and an edge intelligence system 1420. ) Additional details of the embodiment are provided. These factors include, for example, where data is stored and processed (e.g., optimized by AI) in networks and clouds, on-device storage, local systems, and peer-to-peer. By changing , we provide a set of systems that adaptively manage “edge” computation, storage, and processing. These elements may facilitate dynamic definition by users, such as developers, operators, or hosts of platform 102, of a given application anywhere in the world and especially in areas of the ocean where connectivity may be constrained. The purpose is to construct an “edge”. For example, in environments where data connections are slow or unreliable (e.g., if the facility does not have good access to a cellular network (e.g., due to remoteness on Earth), shielding or interference (e.g., due to the density of network-enabled systems), , thick metal hulls of container ships, thick metal container walls, underwater or underground locations, or the presence of large metal objects (e.g. bolts, hulls, containers, cranes, stacked raw materials, etc.) interfere with networking performance), and /or for congestion (e.g., when there are many devices seeking access to limited networking facilities), edge computing capabilities may be deployed on an environment's local area network, on a peer-to-peer network of devices, or on a local value chain entity ( 652). If a strong data connection is available (e.g., a good backhaul facility exists), the edge computing capability may, for example, store frequently used data at that location. It can be deployed in the network for caching to improve input/output performance, reduce latency, etc. Accordingly, adaptive definitions and specifications where edge computing operations are enabled are: or, optionally, automatically determined between fleets or deployed to a geographic area, for example, by an expert system or automated system, which may be based on detected network conditions relative to the environment.In embodiments, edge intelligence 1420 may Enables computation (including where computations take place within the various available networking resources, adaptation of how networking takes place (e.g., by protocol choice), where data storage takes place, etc.), which allows for requirements, prioritization, etc. , and is multi-application aware, such that it takes into account QoS, latency requirements, congestion, and costs, and is understood and prioritized based on recognition of the value of edge computing capabilities across two or more applications.

In an embodiment, digital twin system 1700 may host a floating asset twin 1570 that may be associated with one or more of floating assets 620 . In this example, one or more of the floating asset twins 1570 may simulate how one or more of the floating assets 620 will perform without having to test one or more of the floating assets 620 in the real world. Other examples include visualization of all of the ship's systems, the ship's navigation course, and the ship's floating asset twin (1570), which provides information on everything about the ship, from engine performance to hull integrity, available at a glance throughout the entire life of the ship. Includes functional requirements that include various details of the type of information.

In embodiments, use of floating asset twin 1570 during operation may appear to provide informative visualization of any and all significant components of one or more floating assets 620. The use of floating asset twin 1570 during operation may prove beneficial to perform analysis and improve operation of the structural and functional components of floating asset 620. In another example, use of the floating asset twin 1570 during the operation of one or more of the floating assets 620 may be used to model in situ hydrodynamic and aerodynamic changes to the structure and hull surface of the floating asset 620. In embodiments, floating asset 620 may change the configuration of a cross-section of a particular portion of the hull, change the configuration of a hydrodynamic control surface below the waterline, change the configuration of an aerodynamic control surface above the waterline, and perform certain maneuvers. Systems can be deployed to improve hull stability, such as extending additional buoyancy members from the hull. In this example, artificial intelligence system 1160 uses known travel routes and historical weather patterns to determine a schedule of hull configuration changes to improve fuel efficiency with a simulated hull configuration placed on floating asset twin 1570. can be studied.

In embodiments, the use of floating asset twins 1570 during operation may appear to benefit the operator because the operator can plan more efficient inspection and maintenance of one or more floating assets 620. In embodiments, use of port infrastructure twin 1714 during operation may appear to benefit operators who can plan more efficient inspection and maintenance of one or more physical assets within marine facility 622. This can also lead to a longer lifespan of physical assets, as preventative measures will be taken to avoid damage.

In an embodiment, the use of floating asset twin 1570 during operation creates a visual model of the vessel and its underlying systems, such as engine spaces and pumps, distributed over sources of energy, such as engines, boilers, and batteries. It may appear to provide operators with the ability to continuously record fuel consumption. In this example, the operator may plan for more efficient operation, inspection, and maintenance of one or more floating assets 620. In embodiments, the use of the port infrastructure twin 1714 during operation provides the operator with a visual model of the maritime assets moored in position, in an onshore port, and deployed as a navigation aid, including basic systems such as system power plants. It can be shown to provide the ability to continuously record energy consumption distributed over energy sources such as engines, boilers and batteries. In this example, an operator may plan for more efficient operation, inspection, and maintenance of one or more physical assets within marine facility 622. In embodiments, a digital twin system may include simulation and analysis models that can be developed to obtain optimal fuel consumption for a specific voyage with a specific cargo, by including external factors such as wind, current, and weather conditions. . In embodiments, the digital twin system includes external factors such as weather conditions and other assets monitored by the adaptive intelligence layer 614 to determine optimal energy consumption for specific port activities, such as unloading specific cargo. It may include simulation and analysis models that can be developed to acquire.

In embodiments, the use of the floating asset twin 1570 and the port infrastructure twin 1714 during operation may, among other things, allow the supply chain to span physical assets within one or more floating assets 620 and marine facilities 622, either on land or water. Provides an operator with the ability to visualize the behavior of mechanical systems deployed within one or more floating assets 620 or on physical assets within an offshore facility 622 when a process can be maintained, increased, or decreased based on the progress of others being processed. , can be shown to provide the ability to control and adapt.

In embodiments, the use of floating asset twin 1570 and port infrastructure twin 1714 during operation is intended to provide an optimal point for retrofitting batteries and replacing other switchgear during their service life at sea or on land. It may appear. In embodiments, the use of the floating asset twin 1570 during operation may be used to transform a more powerful, more efficient, or more versatile engine, thruster, or other propulsion system into a more powerful, more efficient, or more versatile engine, thruster, or other propulsion system in a normal maintenance cycle or at an appropriate time for retrofit of a component. It can be shown to provide a basis for change.

In embodiments, the use of the floating asset twin 1570 during operation may be used to adjust the forward bulbous bow of the floating asset 620 to improve efficient flow around the bow of the vessel in various combinations of vessel speed, water activity, and weather. It can be seen that it provides a basis for tuning the schedule for In this example, the front ball bow may adjust its shape based on a predetermined schedule or a modified schedule adjusted by the adaptive intelligence layer 614 for the bow's shape for the most efficient execution.

In embodiments, the use of a floating asset twin 1570 during operation may appear to provide an optimal point during a voyage to perform hull cleaning, maintenance or painting or to perform propeller cleaning, maintenance or replacement. . In an embodiment, the use of floating asset twin 1570 during operation can be used to schedule when hull or propeller cleaning is needed and where along the journey contributes the greatest need for cleaning the system, and simulations using floating asset twin 1570 As such, it may appear to provide a basis for determining whether such maintenance is warranted or whether routing the floating asset 620 in a different corridor may result in a lower maintenance burden.

In embodiments, the use of floating asset twin 1570 during operation will be shown to provide detailed simulation and visualization of optimal points during a voyage to perform hull cleaning, maintenance or painting or to perform propeller cleaning, maintenance or replacement. You can. In an embodiment, the use of floating asset twin 1570 during operation can be used to schedule when hull or propeller cleaning is needed and where along the journey contributes the greatest need for cleaning the system, and simulations using floating asset twin 1570 As such, it may appear to provide a basis for determining whether such maintenance is warranted or whether routing the floating asset 620 in a different corridor may result in a lower maintenance burden.

In embodiments, the use of floating asset twin 1570 during operation may be shown to provide detailed simulation and visualization of the performance of one or more vessels or floating assets 620 at a detailed level, allowing users to improve training and safety. The effects of design choices and changes to one or more vessels or floating assets 620 can be viewed when simulating past voyages, predicted voyages, and previous voyages modified to further simulate the activities encountered. In embodiments, the use of floating asset twins 1570 during operation can monitor the performance of multiple vessels or floating assets 620 at a detailed level such that a user can use the digital twin to benchmark performance against other vessels or marine assets. It can be presented to provide detailed simulations and visualizations, and these comparisons can be used to simulate past voyages, predicted voyages, and previous voyages modified to further simulate the activities encountered to improve training and safety. there is.

In embodiments, use of floating asset twin 1570 may provide vessel owners with the ability to visualize vessels and subsystems (and various other marine assets), qualify and analyze operational data, optimize vessel performance, improve internal and external communications, increase It can be shown to provide tools for safe handling and safe disassembly of certain levels of autonomy.

In embodiments, use of floating asset twin 1570 provides equipment manufacturers with tools to facilitate system integration, demonstrate technical performance, perform system quality assurance, and facilitate additional services for monitoring and maintenance. It may appear that it is provided.

In embodiments, the use of the floating asset twin 1570 and the port infrastructure twin 1714 may be used to control the floating assets, whether ships, barges, or other floating assets, and port infrastructure - moored navigation aids, in unloading and loading conditions. of cargo, and even personnel moving throughout the port infrastructure to ensure its operation - a structured framework that can be set up as an application to generate required reporting and supply live information from each maritime asset. It may appear that work is provided to an institution. In many instances, the use of floating asset twins (1570) and port infrastructure twins (1714) will allow for higher quality response to critical issues without imposing additional burden or cognitive load on crews already ensuring the operation of various maritime assets. This may appear to warrant reporting. In many instances, the use of floating asset twins (1570) and port infrastructure twins (1714) has become the basis for commerce to support maritime activities without imposing additional burdens or cognitive loads on crews already ensuring the operation of various maritime assets. It can be seen to ensure higher quality reporting on legal and regulatory issues by providing a time-stamped ledger of the contracts and activities paired with them.

In embodiments, the use of floating asset twin 1570 and port infrastructure twin 1714 provides universities, colleges, and municipalities with knowledge exchange to increase system understanding and enhance research and development and education in various technology areas. It can be seen as providing a platform that facilitates. In these examples, the use of Floating Asset Twin 1570 and Port Infrastructure Twin 1714 increases a candidate's understanding of an entire vessel or a specific marine asset and all of the assets, including that asset, floating and infrastructure assets ( It could appear to provide a maritime academy platform for training that can train these people in terms of understanding the system to see the integrated results of actions taken when influencing (or part of) the system. In this example, system understanding can be seen to be improved because it is visualized and interpreted from the floating asset twin 1570 and the port infrastructure twin 1714, which both include suggestions from the adaptive intelligence layer 614. The integrated outcome of the actions taken can be how activities at the asset level, the fleet at the asset level, the infrastructure level, and the fleet affect the profitability of the fleet through a combination of improving revenues and reducing costs. This is because it can be shown at a business level to show whether it is possible.

In an embodiment, the information technology system that includes the value chain network management platform 604 includes a maritime fleet management application 880 associated with one or more maritime assets, such as one or more floating assets 620 or assets within a maritime facility 622. You may have the same asset management application 814. In an embodiment, the data handling layer 608 of the management platform 604 may process data, e.g., in the data storage layer 624, collected for any of the value chain entities 652 that include one or more maritime assets. Sources and other inputs, such as from monitoring layer 614. In an embodiment, the data source includes information used to populate a training set based on a set of marine activities of one or more of the marine assets, one of design results, parameters, and data from one or more data handling layers 608. is associated with one or more maritime assets. In embodiments, an artificial intelligence system, such as adaptive intelligence layer 614, may be configured to learn based on one or more of the training sets obtained from data sources from one or more data handling layers 608. In doing so, the artificial intelligence system may simulate one or more design attributes of one or more of the marine assets. The artificial intelligence system may also generate one or more sets of design recommendations based on training sets collected from data sources. In an embodiment, the digital twin system 1700 within the value chain network management platform 604 configures one of the maritime assets to include details generated by an artificial intelligence system of one or more design attributes in combination with one or more sets of design recommendations. It may provide visualization of one or more digital twins of one or more maritime assets.

In embodiments, maritime assets may include one or more container ships. In embodiments, marine assets include one or more barges. In embodiments, marine assets include one or more components of port infrastructure installed on or adjacent to land. In embodiments, maritime assets include one or more moored navigation units deployed on the water. In an embodiment, the marine assets include a vessel and the marine activities include the forward speed of the vessel relative to water and weather conditions based on parameters associated with the energy consumption of a propulsion unit on the vessel.

In an embodiment, the information technology system includes a set of intelligent systems to automatically populate a digital twin of maritime value chain network entities based on data collected by the value chain network management platform 604. In embodiments, a maritime value chain network entity is associated with one or more of the real-world shipyards, and a digital twin may be configured to represent one or more of the real-world shipyards. In embodiments, a maritime value chain network entity may be associated with a real-world maritime port and a digital twin may be configured to represent one or more of the real-world maritime ports. In embodiments, a maritime value chain network entity is associated with one or more container vessels, and a digital twin may be configured to represent one or more container vessels. In embodiments, a maritime value chain network entity may be associated with one or more barges and a digital twin may be configured to represent one or more barges.

In an embodiment, a marine value chain network entity is associated with one or more event investigations 7700, and the digital twin is capable of acting and interacting with other assets during a timeline associated with one or more of the event investigations 7700 so that the marine value chain network entity can interact with other assets. It may be configured to at least partially represent a network entity. In an embodiment, the marine value chain network entity is associated with one or more legal proceedings 7702, and the digital twin is capable of acting and interacting with other assets during the timeline associated with the one or more legal proceedings 7702 so that the marine value chain network entity It may be configured to at least partially represent an entity. In an embodiment, the data collected by the value chain network management platform relates to loss reporting 7704, and the digital twin of the maritime value chain network entity relates to loss reporting 7704 based on the data collected by the value chain network management platform. ) and is configured to simulate the possibility of loss 7708 associated with it.

In an embodiment, the maritime value chain network entity is a port infrastructure facility, and the data collected by the value chain network management platform correlates data between a set of data collectors for one or more physical items 7710 within the port infrastructure facility. Facilitates identifying theft or misuse of port infrastructure facilities by enabling the digital twin to detail one or more physical items (7710) of port infrastructure facilities for at least one of the port infrastructure facility and operator set (7720). It can be configured.

In an embodiment, the maritime value chain network entity is a container ship moored in port infrastructure installed on or adjacent to land.

In an embodiment, the data collected by the value chain network management platform is based on the container ship with parameters associated with at least forward speed for water and weather conditions and energy consumption of the propulsion unit on the container ship.

In an embodiment, the value chain network management platform 604 includes an asset management application 814 associated with the value chain network management platform and one or more maritime facilities connected to a container vessel.

In embodiments, the asset management application is associated with one or more vessels connected to a barge.

In an embodiment, the maritime value chain network entity is one or more vessels and the digital twin may provide visualization of the navigation course of the one or more vessels. In an embodiment, the maritime value chain network entity is one or more vessels, and the digital twin may provide visualization of engine performance of the one or more vessels. In an embodiment, the maritime value chain network entity is one or more vessels, and the digital twin may provide visualization of the hull integrity of one or more vessels.

In embodiments, a digital twin may provide visualization of a plurality of inspection points 7730 and maintenance history 7732 associated with the inspection points on a maritime value chain network entity. In embodiments, the digital twin may further provide visualization of a plurality of inspection points 7730 on a maritime value chain network entity within geofenced parameters 7740 and maintenance history 7732 associated with these inspection points 7730. You can.

In an embodiment, the digital twin is a visualization of a plurality of inspection points 7730 on a maritime value chain network entity within geofenced parameters 7740 and the maintenance history mardst 832 associated with those inspection points 7730 and of the associated activities. Additional details of the ledger (7750) may be provided.

CONTROL TOWER AND ENTERPRISE MANAGEMENT PLATFORM FOR VALUE CHAIN NETWORK

In embodiments, the control tower may include or interface with an enterprise management platform (or “EMP”). In embodiments, an EMP may be configured to create, integrate, support, and/or operate one or more digital twins. Generally, a digital twin is a digital twin that combines data from multiple data sources into an object, asset, system, device, machine, component, equipment, facility, individual or other entity referred to throughout this disclosure or in documents incorporated by reference herein. Incorporating into models and representations of the core characteristics of entities, these include, for example, but not limited to: machines and their components (e.g. delivery vehicles, forklifts, conveyors, loading machines, cranes, lifts, transporters, trucks) , loading machines, unloading machines, packing machines, picking machines, and others, including robotic systems (e.g., physical robots, collaborative robots, “cobots”), drones, autonomous vehicles, software bots, etc. etc); Shipping process, transport process, maritime process, inspection process, transport process, loading/unloading process, packing/unpacking process, construction process, assembly process, installation process, quality control process, environmental control process (e.g. temperature control , humidity control, pressure control, vibration control, etc.), boundary control processes, port-related processes, software processes (including applications, programs, services, etc.), packing and loading processes, financial processes (e.g. insurance processes, reporting) process, transaction process, etc.), value chain processes such as testing and diagnostic process, security process, safety process, reporting process, asset tracking process, etc.; Mobile phones, tablets, dedicated handheld devices for value chain applications and processes, data collectors (including mobile data collectors), sensor-based devices, watches, glasses, wearables, head-worn devices, clothing-integrated devices, bands, bracelets, neckbands Wearable and portable devices such as wearable devices, AR/VR devices, headphones, etc.; Workers, such as delivery workers, delivery workers, barge workers, dock workers, dock workers, train workers, ship workers, distribution or fulfillment center workers, warehouse workers, fleet drivers, business managers, engineers, floor managers, demand managers, Marketing managers, inventory managers, supply chain managers, freight handlers, inspectors, delivery personnel, environmental control managers, financial asset managers, process supervisors and operators (for any of the processes mentioned in this application), security personnel, safety personnel, etc. ; Suppliers, such as suppliers of all types of goods and related services, component suppliers, ingredient suppliers, material suppliers, manufacturers, etc.; Includes consumers, licensees, business owners, businesses, value-added and other resellers, retailers, end users, distributors, and others who may purchase, license, or otherwise use categories of goods and/or related services. customers who do; Air travel facilities, docks, yards, cranes, roll-on/roll-on, including loading and unloading docks, storage and warehousing facilities, vaults, distribution facilities and fulfillment centers, aircraft, airports, hangars, runways, refueling depots, etc. Off-shore facilities, such as port infrastructure facilities such as ramps, containers, container handling systems, waterways, locks, etc., shipyard facilities, floating assets such as ships, barges, boats, etc., facilities at the point of origin and/or point of destination, and A wide range of operational facilities, such as transport facilities such as container ships, barges, and other floating assets, as well as land-based vehicles and other delivery systems used to transport goods, such as trucks, trains, etc.; Items or factors that take into account demand, including market factors, events, etc. (i.e., demand factors); Items or factors considered for supply (i.e., supply factors), including market factors, weather, availability of components and materials, etc.; Logistics factors such as availability of travel routes, weather, fuel prices, regulatory factors, availability of space such as vehicles, containers, packages, warehouses, fulfillment centers, stalls, etc.; Retailers, including online retailers; routes for transport such as waterways, roads, air travel routes, railways, etc.; Robotic systems including mobile robots, cobots, robotic systems to assist human workers, robotic delivery systems, etc.; Drones, including package delivery, site mapping, monitoring or inspection; Autonomous vehicles, such as those for package delivery; Enterprise resource planning platform, customer relationship management platform, sales and marketing platform, asset management platform, Internet of Things platform, supply chain management platform, platform-as-a-service platform, infrastructure-as-a-service platform, software-based data storage platform, analytics platform, artificial intelligence software platforms such as platforms; etc.

68 is a schematic diagram of an example environment of enterprise management platform 8000. In embodiments, EMP 8000 may be integrated with or accessible to a control tower via an application programming interface (API). In some of these embodiments, EMP 8000 may be a set of microservices accessible to the control tower.

In an embodiment, EMP 8000 includes an enterprise configuration system 8002, a digital twin system 8004, a collaboration suite 8006, an expert agent system 8008, and an intelligent service system 8010. In an embodiment, EMP 8000 includes an API system 8014 that facilitates the transfer of data between EMP 8000 and one or more external systems. In some embodiments, intelligent services system 8010 includes an enterprise data repository 8012 that stores data about an enterprise, whereby enterprise data can be stored in a digital twin system 8004, a collaboration suite 8006, and/or Used by expert agent system 8008. Enterprise data store 8012 may store any of a wide variety of data, such as any data involved in the data pipeline described above and throughout this disclosure and documents incorporated by reference in this application. In embodiments, enterprise data repository 8012 may store data that is being used to update the digital twin in real time or substantially real time. In embodiments, enterprise data repository 8012 may store databases, file systems, folders, files, documents, transient data (e.g., real-time data or substantially real-time data), sensor data, etc.

In embodiments, enterprise configuration system 8002 may provide an interface (e.g., a graphical user interface (GUI) )) is provided. As used in this application, enterprise may refer to a for-profit or non-profit organization, company, governmental agency, non-governmental organization, etc. Although described as on-boarding users, the configuration of the enterprise management platform 8000 for a particular enterprise may include individuals associated with the enterprise, individuals associated with the EMP, and/or (cloud resources, platform-as-a-service, software-as-a-service, multi-tenant data). may be performed by any number of users, including service providers and/or individuals associated with third parties, such as third party hosts of hosted EMPs for enterprises (which may be deployed on resources and/or similar resources) .

In embodiments, an on-boarding user may define the type of enterprise digital twin that can be created by digital twin system 8004 on behalf of the enterprise being on-boarded. In embodiments, on-boarding users may select different types of digital twins to be supported for the enterprise by EMP 8000 via a GUI presented by enterprise configuration system 8002. For example, a user can select different types of role-based digital twins from a menu of digital twin types, where the different types of role-based digital twins include executive digital twins. As another example, a user can select the type of organizational digital twin appropriate for the user's organization, such as from a library of industry-specific or domain-specific organizational templates. In some embodiments, each type of executive digital twin may have a set of predefined states (as referenced herein, such terms include states, entities, relationships, parameters, and other characteristics) that are represented in each executive digital twin. ) and has a predefined level of granularity and/or other characteristics for each state in the set. In some embodiments, the set of states, their respective granularity, and/or other characteristics displayed in the executive digital twin may be customized (e.g., by on-boarding users). In these embodiments, a user may define the different states represented in each type of executive digital twin and/or the granularity for each state displayed in the digital twin. For example, if a company's CEO has a finance background, the CEO may want more financial data to be displayed in the CEO digital twin so that the financial data is displayed at a higher granularity, or the CEO may want to use status information to make decisions. Access basic information about the financial models available to the digital twin, such as the model being used (e.g., a financial forecast or projection) or the model used for augmentation of the state (e.g., highlighting significant deviations from expectations). You may want to do it. In contrast, if the CEO has less financial experience or training, the CEO digital twin can consist of summarized financial data and prompts to obtain CFO input when conditions deviate from normal operating conditions (company and/or industry results). may be generated by an intelligent agent trained on a set of). In this example, the CEO digital twin can be configured to display desired financial data fields at a set of levels of granularity defined by the user (e.g., financial data may include various revenue streams, expense streams, etc.). In another example, the CEO may have a technical background. In this example, the CEO digital twin can be configured to display one or more states related to the company's product and R&D efforts, patent development, and product roadmap at a higher level of granularity. In another example, the COO may be responsible for overseeing the company's product team, marketing team, and HR department. In this example, the COO may want to view marketing-related status, product development-related status, and HR-related status at a lower level of granularity. In this example, the COO digital twin can be configured to show visual indicators indicating whether any of the states are in critical, exceptional, or satisfactory conditions. For example, if employee turnover is very high and employee satisfaction is low, the COO digital twin can indicate that the HR-health is at a critical level. In this configuration, the COO can choose to drill down to HR-Status, where he can view employee turnover rates, hiring rates, and employee satisfaction survey results.

In another example, a COO or CTO digital twin may be configured to represent and assist in the discovery and management of interconnections, relationships, and dependencies between enterprise operations and information technology. For example, a COO digital twin or CFO digital twin can be configured to represent a set of operating entities and workflows (e.g., a flowchart representing a production process, assembly process, logistics process, etc.), where the entities (human workers) , robots, processing equipment, and other assets) to generate and hand off sets of outputs (of various types) to the next set of entities within the workflow for further processing. It is indicated to operate on a set of inputs such as information. These may be represented, for example, in a flow diagram showing each entity and its relationship in the flow to other entities. In embodiments, a role-based digital twin (such as a CIO digital twin) may also include any of the types described throughout this disclosure, including sensors, IoT devices, data collection and monitoring systems, data storage systems, edge and other computation systems. may represent information technology systems, such as representing systems, wired and wireless networking systems, etc. Each information technology component or system is a role-based entity, with associated data such as specifications, configuration parameters and settings, processing capabilities, relationships to other components such as data representation, and networking connectivity to other components or systems. Can be displayed in a digital twin. In embodiments, a role-based digital twin may be configured to indicate which information technology entities are located with wired or close wireless connectivity to which operating entities, and which information technology entities are logically associated with which operating entities. (e.g., where cloud resources, compute resources, artificial intelligence resources, database resources, application resources, or other resources are provisioned to support or interact with the behavioral entity, e.g., in a virtual machine, container, or other logical relationship) and Likewise, it can provide a converged view that displays operational technology entities and information technology entities in relation to each other. In embodiments, the converged view presented in the role-based digital twin may thus indicate location-based and/or logical interconnections between operations and information technology. In embodiments, alerts such as those indicating failure modes, congestion, delays, interruptions in service, poor latency, reduced quality of service, bandwidth constraints, poor performance on key performance indicators, downtime, or other issues are converged. It can be provided as an augmentation or overlay of the information technology and operational digital twin, so that a COO, CTO, CIO or other user can see interconnections between information technology and operational entities that may be contributing to the problem. Other types of problems that can be provided as augmentations or overlays include alerts regarding existing conditions and/or predictions or projections of such conditions, such as by an analytical system or predictive artificial intelligence system, such as an expert agent trained to make such predictions. may include. In an example, if high latency in the control system for a warehouse is slowing down the process of picking and packing goods due to the associated edge compute nodes experiencing congestion on the input data path, users of the role-based digital twin may experience congestion in their actions. You may be alerted to the fact that you are being adversely affected by an Therefore, a converged digital twin of operational technology entities and information technology entities provides insight into how executives can adapt operations and/or information technology to improve outcomes and/or avoid anticipated problems before they become catastrophic failures. can be provided.

In embodiments, a user (e.g., on-boarding user) may connect one or more data sources 8020 to EMP 8000. Examples of data sources 8020 that can be connected to the EMP include a sensor system 8022 (e.g., a set of IoT sensors), a sales database 8024 updated with sales figures in real time, a customer relationship management (CRM) system ( 8026), content marketing platforms (8028), news websites (8048), financial databases that track the costs of a business (8030), surveys (8032) (e.g., customer satisfaction and/or employee satisfaction surveys) survey), organization chart (8034), workflow management system (8036), customer database (1S40) storing customer data, external data feeds (e.g., news feeds, promotional feeds, weather feeds, transaction data, pricing data, market data) etc.), data obtained by spidering, webscraping, or otherwise parsing websites and social media sites, data obtained by crowdsourcing, and/or storing third party data. May include, but is not limited to, data from various third party data sources 8038. Data sources 8020 may include additional or alternative data sources without departing from the scope of this disclosure. Once the user has defined the configuration of each executive digital twin - the configuration includes selected states to be displayed (which may include entities, relationships, and properties), features to be enabled, and/or the desired granularity of each state. The user can then define the data sources 8020 that are fed into each executive digital twin, including any of the data sources within the data pipeline described above. In some embodiments, data from one or more of the data sources may be fused and/or analyzed before being fed into each digital twin.

In some embodiments, on-boarding users may include, among other things, an environmental digital twin, an information technology digital twin, an operational digital twin, an organizational digital twin, a supply chain digital twin, a product digital twin, a facility digital twin, a customer digital twin, a cohort digital twin, etc. You can choose from a variety of types of enterprise digital twins supported for your enterprise, including process digital twins and/or process digital twins. In some of these embodiments, users can create these digital twins and define the data sources used to update the enterprise digital twin. In embodiments, a user may define any physical location to be represented as an environmental digital twin (which may be a digital twin of a facility or other suitable environment). For example, users can define manufacturing facilities (e.g., factories), shipping facilities, warehouses, office buildings, etc. Each facility may be given an identifier, such as a location (which may include a logical and/or virtual location and/or a geographic location) and a name and type description. In embodiments, enterprise configuration system 8002 may assign an identifier to each facility and associate the location of the facility with the identifier. In embodiments, users may define types of objects that may be included in and/or found within the environment. For example, users can specify the types of enterprise resources (e.g., factories, warehouses, or distribution centers equipment and machines, assembly lines, conveyors, vehicles, robots, high-lows, etc., IT systems, workers, etc.) in the environment. , types of products, materials and components that are manufactured in the environment, stored in the environment, moved around, assembled, used as inputs within, produced in the environment, sold from the environment and/or received in the environment, in the environment. It can define the types of sensors/sensor kits and/or data collection, storage and/or processing devices used, the operators and workflows involved, etc. Examples of how environmental and process digital twins are created and updated include U.S. Provisional Application Nos. 62/931,193, filed November 5, 2019, and entitled “Methods and Systems of Value Chain Network Management Platform,” and No. 2, 2020 It can be found in U.S. Provisional Application No. 62/969,153, filed on Mar. 3 and entitled “Methods and Systems of Value Chain Network Management Platform,” the contents of which are incorporated herein by reference.

In an embodiment, enterprise organization system 8002 is configured (in combination with digital twin system 8004) to create an organizational digital twin that represents the organizational structure of an enterprise. In some embodiments, an organizational digital twin may represent individuals/roles occupying management and professional levels of an enterprise. Alternatively, an organizational digital twin may include a workforce digital twin representing the entire workforce of an enterprise, including all of the enterprise's workforce and/or contractors, or defined portions thereof. For example, in an enterprise setting, the workforce may include the logistics workforce, warehouse workforce, distribution workforce, reverse logistics workforce, forwarding workforce, factory operations workforce, plant operations workforce, resource extraction operations workforce, (e.g., the internal network of an industrial enterprise) (to operate) may include network operation workforce, sales workforce, marketing workforce, advertising workforce, retail workforce, R&D workforce, technology workforce, engineering workforce, etc. In another example, in the context of a value chain network, the workforce may include supply chain management workforce, logistics planning workforce, vendor management workforce, etc. In another example, in the context of a market setting, labor may include brokerage labor for the market, transaction execution labor for the market, transaction coordination labor for the market, transaction execution labor for the market, etc. Companies may include additional or alternative labor. In some embodiments, an organizational digital twin may include management-level roles within the workforce. Examples of management-level roles in companies include the CEO role, COO role, CFO role, Counselor role, Board Member role, CTO role, Information Technology Manager role, Chief Information Officer role, Chief Data Officer role, Investor role, Engineering Manager role, This includes project manager roles, operations manager roles, and business development roles. Additionally, management level roles in the workforce may include a plant manager role, a plant operations role, a plant operator role, a power plant manager role, a power plant operations role, a power plant operator role, an equipment service role, and an equipment maintenance operator role. . In the value chain context, management-level roles in the workforce include the Chief Marketer role, Product Development role, Supply Chain Manager role, Customer role, Supplier role, Vendor role, Demand Management role, Marketing Manager role, Sales Manager role, and Service Manager role. , may include a demand forecasting role, a retail manager role, a warehouse manager role, a salesperson role, and a distribution center manager role. In the context of markets, the management-level roles of the workforce include the market maker role, exchange manager role, broker-dealer role, trading role, coordination role, contract counterparty role, exchange rate setting role, market orchestration role, market composition role, and contract composition role. May include roles. It is understood that not all roles defined above apply to specific types of workforce. Additionally, some roles may involve different types of workforce.

In some embodiments, an organizational digital twin may further include data access rules for different departments and/or roles within the organization. For example, the CEO may be granted access to most or all of the organization's data, the CFO may be granted access to financial-related data and may be restricted from viewing R&D data, and the CTO may be granted access to R&D-related data. Granted access and viewing of financial data may be restricted, members of the engineering team may be restricted from accessing financial-related data, and so on. Analytical models, including role-based or identity-based control of the ability to view results, configure inputs, configure or adjust models (e.g., weights, inputs, or processing functions), perform control actions, etc. , similar rules may apply to access to features such as artificial intelligence systems, intelligent agents, etc. In some embodiments, the EMP may utilize an organizational digital twin when determining what level of access a particular individual can be granted and/or whether to deny certain types of access to the individual. In some embodiments, access rights specify the types of data a particular user can access, such as information about each individual listed in the organization's digital twin (e.g., salary, start date, availability, work status, etc.). It can be limited. For example, lower-level employees may not be granted access to sensitive information such as financial data, product strategy, marketing strategy, trade secrets, etc. In some embodiments, specific users may be granted permission to change the access rights of other employees, which may be reflected in the organization's digital twin. For example, specific executives and managers can be granted permission to grant access to members of their respective teams when working on specific projects.

In an embodiment, enterprise organization system 8002 receives an organization chart (“organization chart”) definition of an enterprise and creates an organizational digital twin based on the organization chart definition. In embodiments, the organization chart definition may define the business units/departments of the enterprise, the reporting structure of the enterprise, the various roles within the enterprise/each business unit, and the individuals in each role. In some embodiments, a user may upload an enterprise's organizational chart to EMP 8000 via enterprise organization system 8002. Additionally or alternatively, the user can define the structure of the org chart (e.g., roles, business units, reporting structure) and assign various roles to the names and/or other identifiers of the individuals who fill each role defined in the org chart. It can be filled with In some embodiments, enterprise organization system 8002 may be used to obtain enterprise organizational data, such as enterprise roles, individuals filling roles, salaries of individuals filling roles, enterprise reporting structure, etc., to obtain enterprise organizational data, such as enterprise resource planning system 8044 ) and/or the HR system 8046. In these embodiments, the digital twin system 8004 (discussed below) may use the ERP system 8044 and/or the HR system 8046 to receive data necessary to maintain the organizational digital twin in a real-time or near-real-time manner. You can continue to communicate with.

In embodiments, enterprise organization system 8002 (in collaboration with digital twin system 8004, discussed below) may create an organizational digital twin of an enterprise based on the organization chart definition and the individuals who fill the roles within the organization chart definition. In embodiments, a user may define one or more restrictions, permissions, and/or access rights for individuals represented in the organization digital twin through enterprise configuration system 8002. In embodiments, restrictions may define one or more types of data or features that certain users or groups of users are not permitted to access (either directly or in the digital twin). In embodiments, access rights may define one or more types of data or features that a particular user or group of users can access and the types of access that the user or group of users can access. In embodiments, permissions may define actions a user or group of users may perform on EMP 8000. In embodiments, one or more of the access rights, permissions, and restrictions may be geographically defined and/or temporally limited. For example, some types of data or features may be visible or otherwise accessible only in certain areas (e.g., sensitive data only in corporate offices) or at certain times (e.g., during board meetings). In embodiments, restrictions, permissions, and/or access rights may be set for roles or users themselves. Likewise, defining access rights, permissions, and/or restrictions for a user or group of users may also include defining access rights, permissions, and/or restrictions for roles and/or business units within an enterprise. there is. In embodiments, an organizational digital twin may be deployed to manage rights, permissions, and/or restrictions for an enterprise's users. Additionally, in embodiments, an organizational digital twin may define types of role-based digital twins (and other enterprise digital twins) that various users can access. In some embodiments, an organizational digital twin may display additional or alternative information.

In an embodiment, digital twin system 8004 is configured to create, update, and serve an enterprise's enterprise digital twin. In some embodiments, digital twin system 8004 is configured to create and serve role-based digital twins on behalf of an enterprise, and can transmit role-based digital twins to client devices 8050 (e.g., mobile devices, tablets, personal computers). , laptops, AR/VR-enabled devices, workflow-specific devices or equipment, etc.). As discussed, during the configuration phase, users can define the different types of data and the corresponding data sources, data sets, and characteristics that will be used to create and maintain each type of enterprise digital twin. there is. Initially, the digital twin system 8004 is any underlying data source/database (e.g., SQL database, graph database, relational database, distributed database, etc.) that stores or generates data collected by each enterprise digital twin. Constructs data structures that support each type of enterprise digital twin, including blockchains, distributed ledgers, data feeds, data streams, etc. Once the data structures supporting the digital twin are constructed, the digital twin system 8004 receives data from one or more data sources 8020. In embodiments, digital twin system 8004 may structure and/or store received data in one or more databases. When a particular digital twin is requested (e.g., by a user via client application 8052 or by a software component of EMP 8000), the digital twin system may determine the view represented in the requested digital twin and , can create the requested digital twin based on data from a configured database and/or real-time data received via an API. Digital twin system 8004 may serve the requested digital twin to a requestor (e.g., a client application or backend software component of EMP 8000). After the enterprise digital twin is served, some enterprise digital twins may be subsequently updated with real-time data received through the API system 8014. In embodiments, the API provides the data pipeline with information about the types of data needed for the digital twin so that the data pipeline can be configured (either by a user or by an automated/intelligent system) to handle the data effectively. can do. For example, data pipelines can deliver data over data paths that use appropriate protocols for efficient delivery, or cost-appropriate paths (e.g., low-latency or low-cost paths for data that does not require real-time updates). It may be configured to transmit data through . Accordingly, in some embodiments, the configuration of the digital twin may be configured for low latency, high quality of service, high accuracy, high granularity, high reliability, etc., based on the priority of the mission served by data type, for example. This may include providing input regarding the requirements of In embodiments, an intelligent expert agent (or “intelligent agent” or “expert agent”) may be trained on a training set of configurations of inputs to one or more data pipelines previously configured by the expert, such that the intelligent agent has a similar or can learn to automatically configure APIs for digital twins to provide appropriate inputs to the data pipeline for subsequent digital twins that involve similar or similar workflows for similar roles, identities, industries, and/or domains. . In embodiments, such training of intelligent agents may include learning about specific user interactions, such as learning which users within a role use which types of data at what times and for what purposes, so that data resources Appropriately allocated to support actual user requirements. For example, an automated intelligent agent that manages the configuration of the data pipeline for the COO digital twin could be used by operations executives (e.g., COO users) at the end of each eight-hour workday (e.g., 5:00 p.m.). (afterwards) to check production data for each facility, so mid-shift data updates are delivered via lower cost data resources, but end-of-shift data is highly reliable and It is delivered through a low-latency data path with quality of service. Continuing with this example, the intelligent agent can determine how often production data is updated for the COO digital twin, such that the COO digital twin is updated less frequently in the morning and mid-afternoon, but more frequently at the end of business hours. . In embodiments, the intelligent agent is comprised of business logic that defines the overall strategy (e.g., using low-latency networks versus higher-latency networks and/or how often to update certain types of data within a certain digital twin) and , by being customized based on the preferences and usage by the end-user of the digital twin, so that the overall strategy can be learned from training data sets obtained from experts and/or hard-coded by the developer, with the customization pieces tailored to the end-intended user. Learning can be learned from monitoring the use of the digital twin by (e.g., when she typically checks production data from each facility). It is to be understood that additional or alternative examples of such data prioritization strategies and/or other organization strategies are included in this application. For example, upon receipt of input regarding performance requirements, the artificial intelligence capabilities of the data pipeline integrated with, associated with, or supporting EMP 100 may automatically or under user control, at the appropriate time and place. Techniques may be used to provide appropriate resources, including but not limited to: adaptive coding of data path transmissions between networked data communication nodes; Adaptive filtering, repetition, and amplification of RF/wireless signals (including software-implemented bandpass filtering); Dynamic allocation of use of cellular and other wireless spectrum, adaptive, ad-hoc, cognitive management of wireless mesh network nodes; adaptive data storage; cost-based routing of wireless and wired signals; priority-based routing; Selection of channels and performance-aware protocols for communication; Context-aware allocation of computation resources, serverless computation system, adaptive edge computation system, channel-aware error correction, smart contract implementation network resource allocation; and/or other suitable techniques.

In embodiments, digital twin system 8004 may be further configured to perform simulation and modeling on an enterprise digital twin. In embodiments, digital twin system 8004 is configured to run data simulations and/or environmental simulations using digital twins. For example, a user, through a client device, may instruct digital twin system 8004 to perform a simulation for one or more states and/or workflows displayed in the digital twin. The digital twin system 8004 can run a simulation on the digital twin and display the results of the simulation on the digital twin. In this example, the digital twin may need to simulate at least some of the data used to run the simulation of the environment so that reliable data is present when performing the requested environment simulation. Digital twin system 8004 is discussed in greater detail throughout this disclosure.

In embodiments, collaboration suite 8006 provides a diverse set of collaboration tools that can be utilized by a variety of users in an enterprise. Collaboration tools may include video conferencing tools, “in-twin” collaboration tools, whiteboard tools, presentation tools, word processing tools, spreadsheet tools, etc. In embodiments, “in-twin” collaboration tools allow multiple users to view and collaborate within a digital twin. For example, in embodiments, a collaboration tool may be used to identify the same collaborative entities and events (such as version control objects, comment streams, edit events, and other changes) when they are represented within a digital twin interface and are associated with a digital twin entity. May include in-twin collaboration tools that enable digital twin experiences and collaboration experiences within the interface (e.g., within an AR/VR-enabled user interface, standard GUI, etc.). For example, multiple users can be granted access to view the environmental digital twin of a facility, such as a warehouse or factory, through an in-twin collaboration tool. Once viewing the environmental digital twin, the user can then change one or more features of the environment displayed in the environmental digital twin and instruct the digital twin system to perform a simulation. In this example, the results of the simulation can be presented to the user in the digital twin and automatically populated into a shared document (e.g., a spreadsheet or presentation document). As discussed throughout this disclosure, users may collaborate in additional ways with respect to digital twins. For example, in some embodiments, collaboration suite 8006 may allow a user to call a video conference with another user, where the users see each other and view aspects of a particular digital twin related to the topic of discussion for the conference. can see. In this example, users can see each other, for example, at a representation of the work being discussed, so that the user can see gestures or indications from other users about how the work should be acted upon. In another example, Twin's meeting features can show participants by their location within a view of a set of facilities' environments, so that users can recognize which participants have the closest proximity to the relevant assets with which they are collaborating. there is. In some embodiments, collaboration suite 8006 may interface with third-party applications, thereby allowing data to be imported to and/or from third-party applications. For example, when collaborating on a board presentation, different executives can export data from their respective executive digital twins into a shared presentation file (e.g., a PowerPoint™ file or a Google™ Slides presentation). In another example, a first user (e.g., the CEO of an enterprise) can communicate with a second user (e.g., the CEO of an enterprise) through a first executive digital twin configured for the first user (e.g., the CEO digital twin of the enterprise). You can request specific information (for example, a financial plan for your company) from your CTO. In response, the second user may transfer the requested data to EMP 8000 (e.g., collaboration suite 8006 and/or digital twin) from a second executive digital twin configured for the second user (e.g., CTO). to twin system 8004), which in turn can update the executive digital twin configured for the first user. Additional examples and descriptions of collaboration suite 8006 and basic collaboration tools are discussed throughout this disclosure.

In embodiments, the collaboration suite 8006 may configure a collaboration environment and/or task of a collaboration tool such that different participants within the same collaboration environment and/or workflow experience different views or characteristics of the same digital twin entity and/or workflow. Can be configured to interface with digital twin system 8004 (e.g., independently or under control of digital twin system 8004) to provide role-specific views and other features within the flow. For example, the CFO may collaborate with the COO and CTO on possible replacements for internal systems or pieces of machinery or equipment, where the current system, machine, or equipment and/or potential replacement systems, machines, or equipment are evaluated by visual and other elements. It is expressed as a digital twin by . During collaboration, collaboration suite 8006 may recognize the identities/roles of the CFO, COO, and CTO and automatically construct their respective collaboration views into exemplary digital twins based on those roles. For example, the CFO could be presented with a view augmented with financial data, such as the cost of the item and various possible replacements, the term and terms of the lease agreement, depreciation information, information about the financial impact on productivity, etc. Meanwhile, the collaboration suite 8006 provides the COO with information indicating the relationship of items to operational processes, such as links to other systems involved in the production line, timing information (e.g., scheduled downtime for a facility), etc. can be presented. In this example, the CTO provides performance specifications and capability information for an item and, for example, a variety of possible alternatives for other items represented in the digital twin (physical/mechanical compatibility, data compatibility, software compatibility, and many other forms of technology compatibility). You may be presented with a variety of possible alternatives, including compatibility information indicating the degree of compatibility, reviews and ratings, and other technical information. Each executive user must have each information present in that user's “native language” (e.g., information tailored to each executive's respective expertise and needs) and each executive user in a native language that is comfortable for that user. Views and/or features may be presented, while the group collaborates (in live or asynchronous mode) to raise issues, engage in commentary and dialogue, and perform analysis (including simulations as described in this application). Reaching financially prudent, operationally effective, and technically sound decisions (e.g., regarding the selection and timing of replacement or alternatives such as repair). Accordingly, an integrated role-responsive collaboration environment in the context of a shared enterprise digital twin enables collaboration around digital twin entities and workflows while allowing users to engage with role-responsive views and features. In embodiments, the collaboration suite 8006 and/or other systems of the EMP 8000 (e.g., digital twin system 8004) may interact with (text or metadata derived from data or metadata based on state information or other data) The semantic model of the enterprise taxonomy can be accessed to automatically generate and/or provide information presented in the shared digital twin (such as role-specific augmentation of entities with symbols). In embodiments, the enterprise taxonomy may be learned by EMP 8000 through analysis of data provided by the enterprise or may be manually uploaded by a user (e.g., a configuration user associated with the enterprise). Information in a digital twin allows the same entity (e.g., a piece of equipment) to be referred to by different names by different groups in the enterprise (e.g., “asset” by the finance department and “machine” by the operations team). may be presented with a role-specific understanding of the classification system, such as when attributes of entities or associated workflows are assigned role-specific or group-specific different terms, codes, symbols, etc. In embodiments, collaboration suite 8006 may automatically enable translation of terminology between roles, such as converting a description using the name of an entity or describing an attribute of an entity from one role-specific form to another role-specific form. You can do it. Automatic conversion may also suggest alternative terms (e.g., as “asset/machine” or “code red/emergency”). In embodiments, automated transformation may be performed by a transformation model (e.g., an enterprise-specific transformation model) trained by machine learning or similar techniques, whereby the transformation model can be configured to include role-sensitive entities, workflows, and It can be utilized to provide automated transformation for attribute presentation. In embodiments, the transformation model is an unsupervised learning technology that operates on an enterprise's data to identify associations between different terms used by human experts and/or by different roles and/or groups to describe the same thing. It can be trained using a training data set of transformations generated by . In embodiments, the transformation model may be seeded by an explicit transformation model, or may be achieved by deep learning or similar techniques known to those skilled in the art.

In an embodiment, expert agent system 8008 trains expert agents to perform/recommend actions on behalf of an expert. An expert agent may be a software module that implements and/or utilizes artificial intelligence services to perform/recommend actions on behalf of or on behalf of an expert. In embodiments, an expert agent may, in connection with a defined role, create one or more machine learning models that perform machine learning tasks, including robotic process automation (e.g., throughout this disclosure and/or documents incorporated by reference herein). It may include neural networks, prediction models, classification models, Bayesian models, Gaussian models, decision trees, random forests, etc.), including any of the artificial intelligence systems, expert systems, etc. described throughout. Additionally or alternatively, expert agents may be comprised of artificial intelligence rules that determine actions in relation to defined roles. Artificial intelligence rules may be programmed by a user or may be generated by an expert agent system 8008. The expert agent may run on client device 8050 and/or may be run by or by a system associated with or integrated with EMP 8000. In embodiments, expert agents may be accessed as a service (e.g., via an API), e.g., in a service-oriented architecture, which in embodiments may be integrated with the EMP as a service that is part of a microservices architecture. In embodiments where the expert agent runs at least partially on a client device, EMP 8000 can train an executive agent and serve the trained executive agent to a client application 8052. In embodiments, an expert agent may be implemented as a container (e.g., a Docker container), a virtual machine, a virtualized application, etc., that may run on a client device 8050 or on an EMP 8000. In an embodiment, the expert agent is further configured to collect and report data to the expert agent system 8008 that the expert agent system 8008 uses to train/enhance/reconfigure the expert agent. Many examples of such training are described throughout this disclosure, and many other examples are intended to be included by this disclosure.

In some embodiments, expert agent system 8008 (operating in connection with artificial intelligence service system 8010) may utilize, for example, robotic process automation techniques, machine learning techniques, or techniques incorporated by reference into this disclosure and/or this application. Train expert agents (e.g., executive agents and other expert agents) using other artificial intelligence or expert systems, as described throughout the document, to perform activities that are automated by robotic process automation or other technologies. One or more executive actions can be performed on behalf of each user, such as the responsible officer or another user. In some of these embodiments, client application 8052 may be configured to connect a client device associated with a user (e.g., an executive, an administrative assistant to an executive, a board member, role-based professional, manager, worker, or any other suitable employee or affiliate). 8050) (e.g., a user device such as a tablet, AR and/or VR headset, mobile device, or laptop, embedded device, enterprise server, etc.). In embodiments, client application 8052 may record the user's interactions with client application 8052 and report the interactions to expert agent system 8008. In this embodiment, the client application 8052 can determine any stimulus or input presented to the user, what the user was looking at at the time of the interaction, the type of interaction, the user's role, and whether the interaction was requested by another person. , Characteristics related to the interaction, such as the role of the individual requesting the interaction, context information, status information, workflow information, event information, etc., can be additionally recorded and reported. Expert agent system 8008 may receive interaction data and related characteristics and create, train, configure, and/or update executive agents based thereon. In an embodiment, the interactions may be interactions by a user with an enterprise digital twin (e.g., environmental digital twin, role-based digital twin, process digital twin, etc.). In embodiments, the interaction may involve sensor data (e.g., vibration data, temperature data, pressure data, humidity data, radiation data, electromagnetic radiation data, motion data, etc.) and/or physical entities of the enterprise (e.g., machines). , buildings, shipping containers, etc.), data from various corporate and/or third-party data sources (as described throughout this disclosure and incorporated documents), (characteristics, characteristics, It can be interactions by the user with entity data (such as parameters, settings, configurations, properties, etc.), workflow data (such as timings, decision steps, events, task activities, dependencies, resources, etc.), and many other types of data. . For example, a user may be presented with sensor data from a particular piece of machinery or equipment and, in response, may determine that corrective action should be taken with respect to the machine or piece of equipment. In this example, the expert agent may be trained for conditions that would cause the user to take corrective action, as well as cases where the user does not take corrective action. In this example, an expert agent can learn the circumstances under which corrective action is taken.

In embodiments, expert agent system 8008 may train expert agents based on user interactions with network entities and/or computational entities. For example, expert agent system 8008 can train expert agents to learn how IT professionals diagnose and handle security violations. In this example, an expert agent could be trained to learn the steps taken by an expert to diagnose a security breach, the individuals within the enterprise to whom the security breach is reported, and any actions taken by the expert to resolve the security breach. there is.

In embodiments, the types of actions that expert agents can be trained to perform/recommend include: selection of tools, selection of tasks, selection of dimensions, settings of parameters, configuration of settings, among other possible types of actions. , flagging items for review, providing alerts, providing summary reporting of data, selection of objects, selection of workflows, triggering of workflows, sequencing of processes, sequencing of workflows, processing of workflows. Interruption, selection of data sets, selection of design choices, creation of a set of design choices, identification of failure modes, identification of defects, identification of operating modes, identification of problems, selection of human resources, selection of labor resources, human resources Providing guidance, and providing guidance to workforce resources. In embodiments, expert agents may be trained to perform different types of tasks, such as: determining the architecture for the system, reporting status, reporting events, reporting conditions, and conditions. reporting, determining the model, constructing the model, populating the model, designing the system, designing the process, designing the device, engineering the system, engineering the device. engineering processes, engineering products, maintaining systems, maintaining devices, maintaining processes, maintaining networks, maintaining computational resources, maintaining equipment. , maintaining hardware, repairing systems, repairing devices, repairing processes, repairing networks, repairing computational resources, repairing equipment, repairing hardware, Assembling the system, assembling the device, assembling the process, assembling the network, assembling the computational resources, assembling the equipment, assembling the hardware, setting the price, system physically securing the device, physically securing the process, physically securing the network, physically securing the computational resources, physically securing the equipment, physically securing the hardware. Securing physically, cybersecuring systems, cybersecuring devices, cybersecuring processes, cybersecuring networks, cybersecuring computational resources, cybersecuring equipment -Securing, cyber-securing hardware, detecting threats, detecting defects, tuning systems, tuning devices, tuning processes, tuning networks, computing Tuning resources, tuning equipment, tuning hardware, optimizing systems, optimizing devices, optimizing processes, optimizing networks, optimizing computational resources, equipment optimizing hardware, monitoring systems, monitoring devices, monitoring processes, monitoring networks, monitoring computational resources, monitoring equipment, monitoring hardware Monitoring, configuring systems, configuring devices, configuring processes, configuring networks, configuring computational resources, configuring equipment, and configuring hardware. As discussed, the expert agent may be configured to determine an action and output the action to a client application 8052. Examples of outputs from expert agents may include recommendations, classifications, predictions, control instructions, input selection, protocol selection, communication, alerts, target selection for communication, data storage selection, computation selection, configuration, event detection, prediction, etc. there is. Additionally, in some embodiments, expert agent system 8008 may train expert agents to provide training and/or guidance in addition to or instead of, more precisely, outputting actions. In such embodiments, training and/or guidance may be specific to a particular individual or role, or may be used for different individuals.

In embodiments, expert agent system 8008 is configured to provide benefits to experts who participate in the training of expert agents. In some embodiments, the benefit is a reward provided based on results resulting from a user of an expert agent that is trained based at least in part on actions by the expert user. In some embodiments, the benefit is a reward provided based on the expert agent's productivity. For example, if an expert agent trained by an individual is utilized in conjunction with a set of users within the enterprise (or outside the enterprise), the individual's account may be offered benefits such as cash rewards, stock rewards, gift card rewards, etc. there is. As expert agents become more widely used, the benefits to individuals may increase. In some embodiments, the benefit is a reward provided based on a measure of the expert agent's expertise. For example, individuals with more sought-after/valuable skills may be awarded greater benefits than individuals with less sought-after/valuable skills. In some embodiments, the benefit is a share of revenue or profits generated by work generated by an expert agent, or cost savings resulting therefrom. In some embodiments, benefits are tracked using a distributed ledger (e.g., blockchain) that captures information associated with a set of actions and events involving expert agents. In some of these embodiments, smart contracts may govern the management of rewards to expert users.

In some embodiments, the set of expert agents trained by expert agent system 8008 may be redundantly deployed with at least a portion of the enterprise's workforce, where the expert agents perform different role tasks within the enterprise. In some of these embodiments, an expert agent may be responsible for a set of interactions by members of an enterprise's defined workforce (e.g., physical entities, digital twins, among others) while performing a defined set of defined workforce roles. , sensor data, data streams, computational entities, and/or interactions with network entities). In some embodiments, interactions may be parsed to identify chains of actions and/or chains of inferences performed by the workforce, whereby the chains of actions and/or chains of inferences are used to train expert agents. do. In some embodiments, the interaction may be parsed to identify the type of processing performed by the workforce on the set of information, whereby the type of processing is implemented in the configuration of each expert agent. Examples of workforces include factory operations, plant operations, resource extraction operations, network operations (for example, those responsible for running networks for industrial enterprises), supply chain workforce, logistics planning workforce, vendor management workforce, and intermediation for markets. It may include labor, transaction labor for the market, transaction coordination labor for the market, transaction execution labor for the market, etc.

In some embodiments, expert agent system 8008 and/or client application 8052 may monitor results associated with the user's interactions and enhance training of expert agents based on the results. For example, whenever a user takes a corrective action, expert agent system 8008 can determine the outcome (e.g., whether a particular condition or problem was resolved) and whether the outcome is a positive or negative outcome. Expert agent system 8008 can then retrain the expert agent based on the results. Examples of results include: financial results, operational results, defect results, success results, performance indicator results, output results, consumption results, energy utilization results, resource utilization results, cost results, profit results, revenue results, sales results, and production results. It may contain data about at least one. In such embodiments, expert agent system 8008 may monitor data obtained from various data sources after an action is taken to determine the outcome (e.g., whether sales increase/decrease and by how much, energy usage decreases). /increase and how much so, costs decrease/increase and how much so, revenue increase/decrease and how much so, consumption decrease/increase and how much so, defect conditions are resolved, etc.). Expert agent system 8008 may include the results in a training data set associated with the actions performed by the expert that resulted in the results.

In some embodiments, expert agent system 8008 receives feedback from users regarding each executive agent. For example, in some embodiments, a client application 8052 that utilizes an expert agent may provide an interface through which a user can provide feedback regarding actions output by the expert agent. In an embodiment, the user provides feedback identifying and characterizing any errors made by the expert agent. In some of these embodiments, a report may be generated (e.g., by a client application or EMP 8000) indicating the set of errors encountered by the expert. Reports can be used to reconfigure/retrain enforcement agents. In embodiments, reconfiguring/retraining an executive agent includes removing inputs that are a source of error, reconfiguring a set of nodes of an artificial intelligence system, reconfiguring a set of weights of an artificial intelligence system, and It may include reconfiguring the set of outputs, reconfiguring the processing flow within the artificial intelligence system, and/or augmenting the set of inputs to the artificial intelligence system.

In embodiments, an expert agent may be configured, at least in part, to act dually as an expert for a defined role within the enterprise. In this embodiment, expert agent system 8008 trains expert agents based on a training data set that includes a set of interactions by a particular expert worker while performing its respective role. For example, the set of interactions that can be used to train an enforcement agent includes the interaction of the expert with the enterprise's physical entities, the interaction of the expert with the enterprise's digital twin, and the interaction of sensor data acquired from the expert's sensor systems. It may include interactions of data streams generated by physical entities of experts and enterprises, interactions of computational entities of experts and enterprises, interactions of experts and network entities, etc. In some embodiments, expert agent system 8008 parses the training data set of interactions to identify the expert's chain of reasoning for the set of interactions. In some of these embodiments, the chain of inferences may be parsed to identify the type of worker's inferences that can be used as a basis for constructing/training an expert agent. For example, the inference chain can be a deductive chain of inference, an inductive chain of inference, a prediction chain of inference, a classification chain of inference, an iterative chain of inference, a trial and error chain of inference, a Bayesian chain of inference, a scientific method chain of inference, etc. there is. In some embodiments, an expert agent system parses a training data set of interactions to identify the type of processing performed by the expert in analyzing the set of interactions. For example, types of processing include audio processing in analyzing audible information, tactile or "touch" processing in analyzing physical sensor information, olfactory processing in analyzing chemical sense information, and text processing. text information processing in analyzing motion information, taste processing in analyzing chemical information, mathematical processing in operating mathematically on numerical data, and making execution decisions. It may include executive-manager processing, creative processing when deriving alternative options, and analytical processing when selecting from a set of options.

In embodiments, the expert agent includes an executive agent trained to output actions on behalf of the executive and/or the executive's manager. In this embodiment, expert agents can be trained for executive roles, such that users in executive roles can train executive agents by performing their respective roles. For example, an enforcement agent may be trained to perform actions on behalf of a user in an executive role or to recommend actions to the user. In some of these embodiments, client application 8052 may provide the functionality of enterprise management platform 8000. For example, in some embodiments, users can view executive digital twins and/or use collaboration tools through client application 8052. During use of the client application 8052, executives may escalate issues identified in each executive digital twin to other members of the enterprise. Whenever a user interacts with client application 8052, client application 8052 can monitor the user's actions and report those actions back to expert agent system 8008. Over time, expert agent system 8008 can learn how specific users respond to specific situations. For example, if the user is a CFO and whenever a critical state related to revenue or cost is identified in CFO Digital, and the CFO escalates the critical state to the CEO, the expert agent system 8008 may determine the critical revenue state and the critical cost state. You can learn to automatically report to the CEO. Additional implementations of expert agent system 8008 are further discussed in this disclosure.

In an embodiment, artificial intelligence service system 8010 performs machine learning, artificial intelligence, and analytics tasks on behalf of EMP 8000. In an embodiment, artificial intelligence service system 8010 trains machine learning models used by various systems in EMP 8000 to perform some intelligence tasks, including robotic process automation, prediction, classification, natural language processing, etc. Includes machine learning systems. In embodiments, EMP 8000 includes an artificial intelligence system that performs various AI tasks such as automated decision making, robotic process automation, etc. In an embodiment, EMP 8000 includes an analytics system that performs different analyzes across enterprise data to identify insights into the various states of the enterprise. For example, in embodiments, an analytics system may analyze a company's financial data to determine whether the company is financially stable, in critical condition, or in desirable condition. In embodiments, the analytics system may perform analytics in real time as data is collected from various data sources to update one or more states of the enterprise digital twin. In embodiments, the intelligent system includes a robotic process automation system that learns the behavior of each user and automates one or more tasks on behalf of the user based on the learned behavior. In some of these embodiments, the robotic process automation system may configure expert agents on behalf of the enterprise. A robotic process automation system may construct machine learning models and/or AI logic that operate to output actions in response to given stimuli. In embodiments, a robotic process automation system receives a training data set of interactions by an expert and constructs a machine learning model and/or AI logic based on the training data set. In an embodiment, artificial intelligence service system 8010 includes a natural language processing system that receives text/dialogue, determines the context of the text, and/or generates text in response to a request to generate text. Intelligent services are discussed in more detail throughout this disclosure.

In an embodiment, EMP 8000 includes an enterprise data repository 8012 that stores data on behalf of a customer enterprise. In embodiments, each customer enterprise may have an associated data lake that receives data from various data sources 8020. In some embodiments, EMP 8000 receives data through one or more APIs 8014. For example, in embodiments, an API may be configured to obtain real-time sensor data from one or more sensor systems 8022 in an enterprise. Sensor data can be collected from data lakes associated with the enterprise. The digital twin system 8004 and the artificial intelligence service system 8010 can structure data in a data lake and populate one or more respective enterprise digital twins based on the collected data. In some embodiments, data sources 8020 can be from sensor system 8022, from suitable IoT devices, from local networking devices (e.g., repeaters, switches, mesh network nodes, routers, access points, gateways, etc. from wireless and fixed network resources), from general-purpose networking devices (e.g. computers, laptops, tablets, smartphones, etc.), from smart products, from telemetry systems of machines, equipment, systems and components (e.g. on-board diagnostics) data collected by data collectors (including drones, mobile robots, RFID and other readers, and human-portable collectors), and/or other suitable data; It may include a set of edge devices 8042 that collect, receive, and process data from sources. In some of these embodiments, edge device 8042 may be configured to process sensor data (or other suitable data) collected at the “network edge” of the enterprise. Edge processing of enterprise data includes sensor fusion, data compression, computation, filtering, aggregation, multiplexing, selective conversion, batching, packetization, streaming, summarization, fusion, fragmentation, encoding, decoding, transcoding, copying, and storage. , decompression, syndication, augmentation (e.g., with metadata), content inspection, classification, extraction, transformation, normalization, loading, formatting, error correction, data structuring, and/or many other processing actions. there is. In some embodiments, edge device 8042 operates on the collected data and/or operates based on the content of the collected data and/or in response to network conditions, operating conditions, environmental conditions, workflow conditions, entity state information, data characteristics, etc. It may be configured to adjust the output data stream or feed based on the same contextual information. For example, edge device 8042 may stream granular sensor data that is identified as abnormal without compression, while edge device 8042 may consider the data to be within tolerance of normal conditions or if the data is of high interest. Less granular data that reflects characteristics that suggest it is less likely (e.g., statistical or signal characteristics) may be compressed, summarized, or otherwise conveyed. In this way, edge device 8042 can provide a semi-perceptual data stream. Semi-perception in edge device 8042 can be achieved by using process automation, machine learning, deep learning, or other artificial intelligence techniques as described herein to learn and train machines on sets of results or feedback from users. can be improved by In embodiments, EMP 8000 may store data streams in a data lake and/or update one or more enterprise digital twins with some or all of the received data.

In embodiments, client device 8050 may execute one or more client applications 8052 that interface with EMP 8000. In embodiments, client application 8052 may request and display one or more enterprise digital twins. In some of these embodiments, client application 8052 may display an executive digital twin corresponding to the user's role. For example, if a user is designated as Chief Marketing Manager, EMP 8000 may provide a digital twin of the user's enterprise's CMO. In some of these embodiments, user data stored on EMP 8000 and/or client device 8050 may indicate the user's role and/or the types of enterprise digital twins (and their features) the user may have access to. .

In embodiments, client application 8052 may display the requested executive digital twin and provide one or more options for performing one or more respective actions/operations corresponding to the executive digital twin and the states displayed therein. there is. In embodiments, an action/operation may be "drilling down" to a particular state, e-mailing or otherwise notifying another user of the state or set of states, or transferring the state or set of states into a collaborative environment ( For example, word processing documents, spreadsheets, presentation documents, slide shows, models (e.g. CAD models, 3D models, etc.), reports (e.g. annual reports, quarterly reports, etc.), websites, wikis. , dashboards, collaborative environment locations (e.g., Slack™ locations), workflow applications, etc.), sending requests for actions on one or more states from other users, performing simulations, Adjusting interface elements (e.g., changing size, color, position, brightness, presence/absence of display, etc.) may include one or more of the following. For example, the COO or other operations executive can view the COO digital twin in action or the COO digital twin. Conditions that may be displayed in the COO digital twin may include potential problems with one or more pieces of machinery or equipment (e.g., as observed from analyzing a stream of data from one or more sensors on a piece of robotic equipment, among others). may include notice. When looking at a COO digital twin, a user may want to escalate a problem to the CEO, for example, request input from other executives, and/or instruct an operations manager, such as a warehouse or plant manager, to handle the problem. In this example, a client application displaying the COO digital twin may allow the user to select an option to escalate the issue. In response to the user selecting the “Receive” option, client application 8052 sends a submit request to EMP 8000. EMP 8000 can then determine the appropriate user or users to whom the problem should be reported. In some embodiments, EMP 8000 may determine an enterprise's reporting structure from an organizational digital twin of the enterprise to which the user belongs. In this example, if the operations executive chooses to have the operations manager handle the issue, the user can select the option to share the status with other users. The user can then enter the intended recipient's identifier (e.g., email address, phone number, text address, user name, role description, or other identifier of the intended recipient (e.g., in various workflow environments, collaboration environments, etc.) (including an identifier for the recipient, etc.) can be input, and a message indicating a command to the intended recipient can be input. In response, the EMP 8000 communicates the identified status to the intended recipient. can do.

In another example, client application 8052 may display the CFO digital twin to a user (e.g., an enterprise's CFO). In this example, the CFO may be responsible for preparing quarterly reports as requested by the CEO. In this example, the CFO may need P&L data, historical sales data (e.g., quarterly sales data and/or annual sales data), real-time sales data, projected sales data, and historical cost data (e.g., quarterly cost and/or annual sales data). You can view a set of different financial positions, including costs, projected costs, etc. In this example, the CFO may select status to include in annual reporting, including P&L data, quarterly sales data, and quarterly cost data. In response to the user selection, client application 8052 may send a request to export the selected status to annual reporting. In this example, EMP 8000 may receive the request, identify a document (e.g., an annual report), and include the selected status in the identified document.

In embodiments, client application 8052 allows a user to respond to a specific request (e.g., a request from the CEO to fill out a report) or notification (e.g., a notification that a machine piece requires maintenance). It may contain a monitoring agent that monitors. The monitoring agent may report the user's responses to these prompts to EMP 8000. In response, EMP 8000 may train an enforcement agent (which may include one or more machine learning models) to handle such notifications the next time they arrive. In some embodiments, the monitoring agent may be included in an enforcement agent included in the client application 8052.

69 illustrates an example set of components of a digital twin system 8004. As discussed, digital twin system 8004 creates visual and/or data-based digital twins, including enterprise digital twins, and creates digital twins for clients (e.g., user devices, servers, and/or digital twins). It is configured to serve internal and/or external applications that utilize . In an embodiment, digital twin system 8004 is an infrastructure component of EMP 8000. In an embodiment, digital twin system 8004 is a microservice accessible by EMP 8000 and/or other components of the value chain control tower.

In an embodiment, digital twin system 8004 includes a processing system 8100 that includes one or more processors, a storage system 8120 that includes one or more computer-readable media, and a network (e.g., the Internet, a private network, etc. ) is executed by a computing system (e.g., one or more servers) that may include a network interface 8130 that includes one or more communication units in communication with. In the illustrated example embodiment, processing system 8100 includes digital twin construction system 8102, digital twin I/O system 8104, data structuring system 8106, digital twin creation system 8108, and digital twin perspective. One or more of a builder 8110, a digital twin access controller 8112, a digital twin interaction manager 8114, a digital twin simulation system 8116, and a digital twin notification system 8118 may be implemented. Processing system 8100 may implement additional or alternative components without departing from the scope of this disclosure. In embodiments, the storage system 8120 may be an enterprise data lake 8122, a digital twin data store 8124, a behavioral data store 8126, and/or a distributed data store, such as a blockchain or a set of distributed data storage resources. It can store corporate data like any other data repository. Storage system 8120 may store additional or alternative data stores without departing from the scope of this disclosure. In embodiments, digital twin system 8004 may be configured with other components of EMP 8000, such as enterprise configuration system 8002, collaboration suite 8006, expert agent system 8008, and/or artificial intelligence service system 8010. You can interface with

In an embodiment, the digital twin configuration system 8102 establishes and manages the enterprise digital twin and the enterprise's associated metadata, configures the data structures and data listening threads that operate the enterprise digital twin, and configures access characteristics, processing characteristics, and automation. It is configured to configure the characteristics of the enterprise digital twin, including features, reporting features, etc., each of which (e.g., the role(s) the enterprise digital twin serves, the entities it represents, and the characteristics the enterprise digital twin supports or enables. This can be influenced by the type of enterprise digital twin (based on the workflow, etc.). In an embodiment, digital twin configuration system 8102 receives the types of digital twins to be supported for the enterprise, as well as the different objects, entities, and/or states to be represented in each type of digital twin. For each type of digital twin, digital twin configuration system 8102 may configure one or more data sources and types of data that supply or otherwise support each object, entity, or state represented by each type of digital twin. and make any internal or external software requests (e.g., API calls) to obtain the identified data types or other suitable data acquisition mechanisms, such as webhooks, configured to automatically receive data from internal or external data sources. You can decide. In some embodiments, the digital twin configuration system 8102 may resolve internal and/or external software requests that support the identified data types by analyzing the relationships between different types of data and their granularity corresponding to specific states/entities/objects. decide Additionally or alternatively, the user may define (e.g., via a GUI) data sources and/or software requests and/or other data acquisition mechanisms that support each data type displayed in each digital twin. . In this embodiment, a user can indicate the data sources to be accessed and the type of data to be obtained from each data source. For example, if a user is constructing an enterprise digital twin of a supply chain process, the user may need an inventory management system to obtain inventory levels, various supplier systems to obtain price data for specific items, and various Sensor systems for acquiring sensor data from a point (e.g., manufacturing facility, warehouse facility, etc.), and other suitable systems for other suitable data types, may be identified. In this data definition process, users can associate specific data types and/or data sources to corresponding structural elements (e.g., layouts, spatial elements, processes, or components thereof) of the digital twin. For example, a user may purchase a product by specifying the specific cost of the product (e.g., the cost of a bearing on a compressor, a headlight on a car, a car, or any other suitable product) obtained through an API request to the seller of the product. It can be matched with the digital twin element it represents (for example, a 3D model of the product). In this example, the digital twin of the product can display the cost of the product, and as the price of the product changes, it can also display a representation of the product.

In an embodiment, composition system 8102 creates one or more external keys for each digital twin that collectively associate different data types with structural elements of the digital twin. Accordingly, when a digital twin is created, foreign keys can be utilized to link data obtained from data sources to structural elements of the digital twin. In some embodiments, a configuration user may define an association that is used to generate a set of foreign keys.

In embodiments, digital twin configuration system 8102 determines, defines, and manages the data structures needed to support each type of digital twin, such as a data lake, relational database, SQL database, NOSQL database, graph database, etc. For example, for an environmental digital twin, the digital twin construction system 8102 may have a database (e.g., an ontology of the environment and a graph database defining the objects that exist (or potentially exist) within the environment and the relationships between them. ), whereby the database instantiated contains and/or references the underlying data that powers the environmental digital twin (e.g., sensor data and analytics related thereto, 3D maps, physical asset twins within the environment, etc.) . In some embodiments, a user can define an ontology for each digital twin, such that the ontology defines the types of data represented in the digital twin and the relationships between those data types. Additionally or alternatively, digital twin construction system 8102 may derive an ontology based on the type of digital twin to be constructed.

In some embodiments, different types of enterprise digital twins may be configured according to sets of preference settings, granularity settings, alert settings, taxonomy settings, topology settings, etc. In some embodiments, configuration system 8102 may configure default preferences for different types of enterprise digital twins, including those that are domain-specific, role-specific, industry-specific, workflow-specific, etc. templates), taxonomies (e.g., default taxonomies for different types of enterprise digital twins), and/or topologies (e.g., graph-based topology, tree-based topology, serial topology, flow-based topology, loop-based topology). Default topologies for different types of twins are available, such as base topology, network-based topology, mesh topology, etc. Additionally or alternatively, configuration system 8102 may receive custom preference settings and categories from configuration users. Non-limiting examples of role-specific templates used to construct a role-based digital twin include CEO template, COO template, CFO template, Counselor template, Board member template, CTO template, Chief Marketing Manager template, Information Technology Manager template, and Chief Information Officer template. Administrator template, Chief Data Officer template, Investor template, Customer template, Vendor template, Supplier template, Engineering Manager template, Project Manager template, Operations Manager template, Sales Manager template, Salesperson template, Service Manager template, Maintenance Operator template, and/ Alternatively, you can include business development templates. Similarly, examples of taxonomies used to construct different types of role-based digital twins include CEO taxonomy, COO taxonomy, CFO taxonomy, Counselor taxonomy, Board member taxonomy, CTO taxonomy, and Chief Marketing Officer taxonomy. system, information technology manager taxonomy, chief information officer taxonomy, chief data officer taxonomy, investor taxonomy, customer taxonomy, vendor taxonomy, supplier taxonomy, engineering manager taxonomy, project manager taxonomy, operations manager taxonomy. , may include a sales manager taxonomy, a salesperson taxonomy, a service manager taxonomy, a maintenance operator taxonomy, and/or a business development taxonomy. Each role-specific template may contain data types specific to the kinds of interactions the role may have and specific responses to those interactions, which may be role-based. For example, a CEO template can include data type definitions for supplier information and labor cost information across the entire organization, and interactions with the CEO digital twin, such as drilling down to specific suppliers and/or labor groups within the enterprise. It may include a response to .

In an embodiment, digital twin configuration system 8102 may be configured to store each enterprise digital twin (e.g., role-based digital twin, environmental digital twin, organizational digital twin, process digital twin, etc.) can be configured to configure and instantiate the database that supports each. In embodiments, for each database configuration, digital twin configuration system 8102 may identify and connect any external resources needed to collect data for each respective data type. For each identified external resource, digital twin configuration system 8102 may configure one or more data collection threads to access APIs, SDKs, ports, webhooks, search facilities, database access facilities, and/or other connectivity facilities. You can. For example, certain executive digital twins (e.g., CEO digital twin, CFO digital twin, COO digital twin, and CMO digital twin) may each require data derived from and/or obtained from the enterprise's CRM 8026. . In this example, the digital twin configuration system 8102 may provide one or more data accesses to the APIs, SDKs, ports, webhooks, search facilities, database access facilities, and/or other connectivity facilities of the enterprise's CRM 8026 on behalf of the enterprise. You can configure a collection thread and obtain any necessary security credentials to access the API. In another example, to collect data from one or more edge devices 8042 of an enterprise, configuration system 8102 may initiate a process to authorize access to the edge devices 8042 of the enterprise to the API of EMP 8000. Thus, the edge device 8042 can provide digital twin data to the EMP 8000.

In embodiments, the digital twin I/O system 8104 supports data sources (e.g., users, sensor systems, internal and/or external databases, software platforms (e.g., CRM, ERP, CRM, workflow management systems). ), surveys, customers, etc.). In some embodiments, digital twin I/O system 8104 (or other suitable component) allows users affiliated with an enterprise to upload various types of data that can be utilized to create an enterprise digital twin of the enterprise. A graphical user interface may be provided. For example, in providing data to support an environmental digital twin, the digital twin I/O system 8104 allows users to access 3D scans, static and video images, LIDAR scans, structured light scans, blueprints, 3D Floor plan, object type (e.g. product, sensor, machine, furniture, etc.), object characteristics (e.g. material, physical properties, description, price, etc.), output type (e.g. sensor unit), architectural drawing , CAD documents, equipment specifications, etc. can be uploaded. In embodiments, the digital twin I/O system 8104 processes data streams (e.g., publicly available data streams such as RSS feeds, news streams, event streams, log streams, sensor system streams, etc.) on behalf of an enterprise. You can subscribe or receive automatically in other ways. Additionally or alternatively, the digital twin system I/O system 8104 may be used in facilities (e.g., manufacturing facilities, shipping facilities, warehouse facilities, logistics facilities, retail facilities, distribution facilities, agricultural facilities, resource extraction facilities, computing facilities). a sensor system 8022 having sensors that sense data from facilities, transportation facilities, infrastructure facilities, networking facilities, data center facilities, etc.) and/or other physical entities of the enterprise, and a sales database that is updated with sales figures in real time ( 8024), CRM systems (8026), content marketing platforms (8028), financial databases (8030), surveys (8032), organizational charts (8034), workflow management systems (8036), third-party data sources (8038), a customer database 8040 to store customer data, and/or a third-party data source 8038 to store third-party data, and an edge device 8042 to report data related to physical assets (e.g., an enterprise Periodically collects data from connected data sources 8020, such as smart machines/manufacturing equipment, sensor kits, autonomous vehicles, wearable devices, etc.), enterprise resource planning systems 8044, HR systems 8046, content management systems 8026, etc. Can query and/or receive. In embodiments, digital twin I/O system 8104 may utilize a set of web crawlers to obtain data. In an embodiment, digital twin I/O system 8104 may include a listening thread that listens for new data from each data source. In an embodiment, digital twin I/O system 8104 may consist of a set of webhooks that receive data from each set of data sources. In this embodiment, digital twin I/O system 8104 may receive data pushed from an external data source, such as real-time data.

In some embodiments, digital twin I/O system 8104 is configured to serve acquired data to a client device 8050 or an instance of an enterprise digital twin (used to populate the digital twin) executed by EMP 8000. do. In an embodiment, the digital twin I/O system 8104 receives a data stream that supplies a received data stream received and/or collected on behalf of an enterprise, and transmits at least a portion of the stream to a data lake 8122 associated with the enterprise. Save it to In embodiments, data streaming into data lake 8122 may be structured and stored in one or more databases stored in digital twin data store 8124.

In an embodiment, data structuring system 8106 is configured to process and structure data into a format that can be consumed by an enterprise digital twin. In embodiments, processing by data structuring system 8106 may include compression, computation, filtering, aggregation, multiplexing, selective conversion, bundling, packetization, streaming, summarization, fusion, fragmentation, encoding, decoding, transcoding, encryption, Decryption, replication, deduplication, normalization, cleansing, identification, copying, storage, decompression, syndication, augmentation (e.g. by metadata), content inspection, classification, extraction, transformation, loading, formatting, error correction, It may include data structuring, and/or many other processing actions. In embodiments, data structuring system 8106 may utilize ETL (extract, transform, load) tooling, data streaming, and other data integration tools to structure various types of digital twin data. In embodiments, data structuring system 8106 structures data according to a digital twin data model that may be defined by digital twin organization system 8102 and/or a user. In embodiments, a digital twin data model may refer to an abstract model that organizes elements of enterprise-related data and standardizes how those elements relate to each other and the characteristics of the digital twin entity. For example, a digital twin data model of an environment that contains a vehicle (e.g., a vehicle assembly facility or an environment in which the vehicle operates) may require data elements representing the vehicle to be sub-elements or attributes of the vehicle (e.g., the vehicle's color, the vehicle's dimensions). , the vehicle's engine, the vehicle's engine parts, the vehicle's owner, the vehicle's performance specifications, etc.). In this example, the digital twin model component can define how physical properties are tied to each physical location on the vehicle. In embodiments, a digital twin data model may define a formalization of objects and relationships found in a specific application domain. For example, a digital twin data model can represent the customers, products, and orders found in a manufacturing enterprise and how they relate to each other within various digital twins. In another example, a digital twin data model may define a set of concepts (e.g., entities, properties, relationships, tables, etc.) that are used to define this formatting of data or metadata within an environment. For example, a digital twin data model used in connection with a banking application may then be defined using an entity-relationship data model and how the entity-relationship data model relates to various executive digital twin views.

In an embodiment, digital twin creation system 8108 serves an enterprise digital twin on behalf of an enterprise. In some cases, digital twin creation system 8108 receives (e.g., via an API) a request for a particular type of digital twin from a client application 8052 executed by client device 8050. Additionally or alternatively, digital twin creation system 8108 receives requests for specific types of digital twins from components of EMP 8000 (e.g., digital twin simulation system 8116). The request may indicate the enterprise, the type of digital twin, the user (whose access rights may be verified or determined by the digital access controller 8112), and/or the user's role. In some embodiments, the digital twin creation system 8108 can define data structures, the grain of the data, response patterns to specific inputs, animation sequences to illustrate behavior, and display for smaller displays (such as mobile phones). Aggregation methods, immersive data interaction systems, security constraints on data observation, observation interaction speed (frame rate), nature of light sources (real or simulated continuous), multiple user participation protocols, network bandwidth constraints, metadata. , the digital twin configuration, as well as information about hooks for ontology and data feeds, can be determined and provided to the client device 8050 (or request component). This information can be used by the client to create a digital twin on an end-user device (e.g., an immersive device such as an AR or VR device, tablet, personal computer, mobile, etc.). In an embodiment, the digital twin creation system 8108 can be configured to include device-sensitive perspectives (e.g., via a digital twin perspective builder 8110, such as delivering them in an appropriate format based on the type of end user device). ) an appropriate view of the requested digital twin and the length of any data limitations, interaction limitations, depth of data limitations, usage limitations, visibility limitations that the user may have (e.g., via access controller 8112). You can decide. In response to determining the viewpoint and data constraints, digital twin creation system 8108 may create the requested digital twin. In some embodiments, creating a requested digital twin involves identifying appropriate data structures, given the viewpoint, and obtaining data that parameterizes the digital twin, as well as any additional metadata that is served with the enterprise digital twin. It may include:

In embodiments, digital twin creation system 8108 may deliver an enterprise digital twin to a requesting client application 8052 (or requesting component). In embodiments, the digital twin creation system 8108 (or other appropriate component) generates real-time data (or real-time data) as it is received and potentially analyzed, extrapolated, derived, predicted, and/or simulated by the EMP 8000. The digital twin served can be continuously updated with data derived from the data.

In some embodiments, the digital twin creation system 8108 (in combination with the digital twin I/O system 8104) includes relational databases, API interfaces, direct sensor input, human-generated input, Hadoop file storage, operations in the environment, and Data streams can be obtained from traditional data sources such as graph databases, telemetry data sources, onboard diagnostic systems, blockchains, distributed ledgers, distributed data sources, feeds, streams, and many other sources, which serve as the basis for reporting tooling. In embodiments, the digital twin creation system 8108 may be used to provide information such as the layout and 3D object characteristics of entities within a facility, a geospatial information system, the hierarchical design of a system of account, and/or the logical relationships of entities and actions in a workflow. A data stream associated with the structural aspect of the data can be obtained. In embodiments, a data stream may include a data stream that includes primary data (e.g., sensor data, sales data, survey data, etc.) and a metadata stream associated with characteristics of the data. For example, metadata associated with a physical facility or other entity may include the type and hierarchy of data being managed, while primary data may contain instances of objects that fall within each hierarchy. The hierarchies at which metadata can be tracked and/or created include, for example, the entire facility, component systems and assets within the facility (equipment, network entities, workforce entities, assets, etc.), subcomponents and subsystems, and optionally down. Metadata for the properties, parameters, or representations of additional subcomponents and subsystems at lower levels of granularity (for example, a ball bearing in a rotating axle assembly of a fan that is part of a motor assembly that drives an assembly line) from a location in a warehouse. Can contain data. In embodiments, hierarchies may include logical or operational hierarchies, such as reporting structures, such as from COO to VP of Operations, from Distribution Manager to Warehouse Manager, or from Shift Manager to Warehouse Workers, among other examples. there is. In an embodiment, the hierarchy extends from the overall process to its subcomponents and decision points, e.g., collection of input materials and components, positioning of the operator, a series of assembly steps, inspection of the output, and post-assembly locations. It may include a workflow or process flow layer, spanning the entire assembly process with a sub-layer of delivery.

In embodiments, the digital twin perspective builder 8110 may use metadata, artificial intelligence, heuristic methods, 3D rendering algorithms, and/or other information in the digital twin creation system 8108 to generate a definition of the information required for creation of the digital twin. Utilize data processing technology. In some embodiments, different related data sets are linked to a digital twin (e.g., executive digital twin, environmental digital twin, etc.) at an appropriate level of granularity, thereby linking structural aspects of the data (e.g., systems of account, sensors, etc.). reads, sales data, etc.) to become part of the data analysis process. One aspect of creating a perspective capability is that a user can change the structural view or granularity of the data, potentially anticipating future events or changes to the structure to guide control of the business area in question. In embodiments, the term "grain of data" may refer to the base unit of a type of data, such as a single line of data, a single aggregated line of data, a single byte of data, a single file, a single instance, etc. You can. Examples of a “grain of data” include a detailed record of a single sale, a single block in a blockchain on a distributed ledger, a single event in an event log, a single vibration reading from a vibration sensor, or similar single or atomic data units. It can be included. Grain, or atomicity, can impose constraints on how data can be combined or processed to form different outputs. For example, if some element of the data is captured at only one level per day, unless it is derived from a daily representation (e.g., using inference techniques and/or statistical models), this may be the case for a single day (or several levels). It can only be decomposed into counts of days) and not hours or minutes. Similarly, if data is only available at the aggregate business unit level, it can only be decomposed to the level of individual employees, for example by averaging, modeling, or derivation functions. In general, role-based and other enterprise digital twins can often benefit from finer levels of data because the aggregation and other processing steps are inherently dynamic and/or involved in dynamic processes and/or real-time decision-making. This is because it can generate output. Note that different types of digital twins may have different “sizes” of data grains. For example, the grains of data that feed the CEO digital twin may be at a higher level of granularity than the grains of data that feed the COO digital twin. However, in some embodiments, the CEO may be able to drill down into the state of the CEO digital twin, and the granularity for the selected state may be increased.

In an embodiment, viewpoint builder 8110 adds relevant viewpoints to the data underlying the digital twin provided to digital twin creation system 8108. In embodiments, a “perspective” may be an adjustment to the ontology of a particular digital twin (e.g., a role-based digital twin) that provides an appropriate ontological view of the underlying data with the correct type at an appropriate level of granularity, aggregation thereof, simplification thereof, and/or may refer to adding details thereto. For example, a CEO digital twin can be linked to market data in fuzzy data and display the potential impact of market forces on a simulated digital twin environment for different scenarios. In another example, in a CFO-level digital twin, an account's internal financial systems can track revenue generation, cost allocation, and structural aspects of the business (e.g., factory floors, warehouses, distribution centers, logistics facilities, office buildings, retail locations, The layout of a container ship, etc. can be assigned across the physical structure of a digital twin, providing the ability to understand the relationships between them. Continuing with this example, the CTO digital twin can include data overlays with current market information about new technologies and the links between them. In this example, the CTO digital twin establishes a link between the impact of a changing technology platform and external information that can be used to improve the facility. These different perspectives generated by the perspective builder 8110 can be combined with the digital twin simulation system 8116 to provide a relevant simulation of how scenario-based future states may be handled by the facility, and the digital twin simulation system 8116 Provides recommendations on how to structurally improve the digital twin representation facilities to meet future state needs, responses to specific changes in the digital twin environment, or changes in information related to digital twin simulation elements. In embodiments, perspective builder 8110 may build perspectives that display intersections or overlays of information technology states and operational states and entities, which can be used to represent information technology and operational technologies within a wide range of industries and domains of operations. May facilitate recognition of opportunities and/or problems involving interaction and convergence. In a further embodiment, perspective builder 8110 may build perspectives that allow different roles to interact with the same digital twin while maintaining different perspectives on operational states and entities, which allows these different roles to have role-specific perspectives. Allows for having meaningful interactions while maintaining In an embodiment, perspective builder 8110 builds a perspective on the digital twin by providing each different user/role with a respective schematic represented as in the digital twin, where these schematics are at a level relevant to that particular user's role. Includes information and structure in. These user-specific diagrams are then linked to underlying data to provide a role-based digital twin experience.

In an embodiment, the digital twin access controller 8112 notifies the creation system 8108 of specific constraints around the roles of the users who can view the digital twin, as well as constraining the view of data or other features specific to each user role. Provides a dynamically adaptable digital twin that can be adapted to be turned on or off. For example, sensitive payroll data may be obfuscated to most management employees when viewing an organization's digital twin, but the CEO may be granted access to view payroll information directly. In embodiments, digital twin access controller 8112 may receive a user identifier and one or more data types. In response, digital twin access controller 8112 may determine whether the user indicated by the user identifier has access to one or more data types or other features. In some of these embodiments, a digital twin access controller may search for a user in the user's enterprise organizational digital twin and determine the user's permissions and restrictions based thereon. Alternatively, a user's permissions and restrictions may be displayed in a user database. In embodiments, an organizational digital twin may be, as noted above, a hierarchical organizational chart, a graph of the organization with nodes representing organizational entities (e.g., work groups, roles, assets, and people), and relationships (e.g., automatically by parsing available data sources to automatically construct a representation of the organization, such as links or associations representing reporting relationships, lines of authority, group affiliations, etc., and data or metadata representing other properties of entities and relationships. can be created.

In an embodiment, the digital twin interaction manager 8114 may be responsible for linking a structural view of the data in the enterprise digital twin (e.g., as displayed/represented by the client application 8052) and the underlying data streams and data sources. Manage relationships. In embodiments, this interactive layer makes the digital twin a window into the underlying data stream through the lens of the structure of the data. In an embodiment, digital twin interaction manager 8114 may be responsible for processing data being fed to an instance of an enterprise digital twin (e.g., an environmental digital twin or an executive digital twin) while the instance is being executed by a client application 8052. Determine the type, or characteristics, of the human interface for establishing this interaction. In other words, the digital twin interaction manager 8114 determines and serves data about the digital twin in use. In an embodiment, the digital twin interaction manager 8114 has specific user interactions and controls that govern the relationship between the user interface and the role-based digital twin. Additionally, in embodiments, such role-based digital twin interactions may be interactions with shared digital twins with different roles interacting seamlessly. In embodiments, digital twin interaction manager 8114 may process raw data received from a data source into a digital twin or digital twin I/O system 8104, or role-based human interaction with digital twin I/O system 8104. Provides a combination of actions. For example, sensor readings of temperature throughout the environment may be fed directly to the environment's execution environment digital twin via the digital twin I/O system 8104, which may then communicate with the environment digital twin to adjust the temperature settings of the environment. In response to human interaction, digital twin interaction manager 8114 may issue control signals to temperature controllers within the environment to increase or decrease temperature.

In embodiments, digital twin interaction manager 8114 obtains data and/or instructions derived by other components of EMP 8000. For example, the CEO digital twin may display analytical data obtained from the artificial intelligence service system 8010 derived from incoming financial data, marketing data, operational data, and sensor data. In this example, digital twin interaction manager 8114 may receive a request from a user to drill down into analytics data, and in response, digital twin interaction manager 8114 may generate financial data, marketing data, etc. from which the analytics data is derived. Data and/or sensor data may be obtained. In another example, digital twin interaction manager 8114 can receive simulated cost data from digital twin simulation system 8116 to convey revenue/cost for different asset maintenance schedules, and the data simulated by is derived using the company's past maintenance data and past sensor data collected by sensors within the company's facilities. In this example, digital twin interaction manager 8114 may receive a request for a different maintenance schedule from a client device representing an executive digital twin (e.g., CFO digital twin, CTO digital twin, or CEO digital twin). and can initiate a simulation for each of the different maintenance schedules. Digital twin interaction manager 8114 may then serve the results of the simulation to the requesting client application.

In embodiments, digital twin interaction manager 8114 may manage one or more workflows performed through the executive digital twin. For example, EMP 8000 may store a set of execution workflows, where each execution workflow corresponds to a role within the enterprise and includes one or more stages. In embodiments, digital twin interaction manager 8114 may receive requests to execute workflows. The request may indicate workflow and user identifiers. In response, digital twin interaction manager 8114 can retrieve the requested workflow and provide specific commands, and/or data, including role-based interactions, to client device 8052.

In an embodiment, digital twin simulation system 8116 receives a request to run a simulation using one or more digital twins. In embodiments, the request may indicate a set of parameters to be changed and/or one or more simulation results to be output. In embodiments, digital twin simulation system 8116 may request one or more digital twins from digital twin creation system 8108 and change different sets of parameters for simulation. In embodiments, digital twin simulation system 8116 may construct new digital twins and new data streams within existing digital twins. In embodiments, digital twin simulation system 8116 may perform environmental simulation and/or data simulation. Environmental simulation focuses on simulation of digital twin ontologies rather than underlying data streams. In an embodiment, digital twin simulation system 8116 generates simulated data streams appropriate for each digital twin environment. This simulation allows for real-world simulations of how the digital twin will respond to specific events, such as changes in the cost of goods supplied or changes in demand for the facility's output.

In embodiments, the digital twin simulation system 8116 serves, in some cases, to develop a framework within which the data and responses of a digital twin can be simulated in response to different situational or contextual inputs/stimuli, including specific response patterns ( Implements a set of models containing (for example, physical and mathematical predictions, logical expressions, or process diagrams). In embodiments, digital twin simulation system 8116 may include or utilize a computerized model builder to construct predicted future states of either data and/or the digital twin's response to input data. In some embodiments, a library of computerized models may be used to represent various types of scientific, economic, statistical, psychological, sociological, econometric, engineering, mathematical, Physical, chemical, biological, architectural, computational, or other models, formulas, functions, processes, algorithms, etc. (collectively referred to in this application as “behavioral models” or “models” except where the context indicates otherwise) 8126 may be obtained from a behavior model data store 8126 that stores one or more models defining one or more behaviors of an entity. In embodiments, a Value Chain Network Data Object is an object-oriented data model that defines the classes, objects, properties, parameters, and other characteristics of a set of data objects (such as those associated with Value Chain Network entities and applications) handled by the platform. It can be provided accordingly. A computerized digital twin model computes the model's output based on available inputs, allowing users to interact with and view key features of the simulated environment while observing how the entire system responds to specific changes in the environment. Build an environment. For example, a digital twin simulation can display how a set of objects stacked in a container will respond to tilting the container, where the behavior of the objects can be combined with a mechanical engineering model of the stacked object, including structural features, weight distribution, etc. /or based on an architectural model. This involves evaluating the probability and/or impact of various failure modes such as failure, leakage, etc., in response to seismic events, road conditions, weather conditions, wave action, etc., as well as the chain of events and other objects in the simulated environment. Can help simulate responses. This may, for example, allow a user to identify events and consequences that occur as a result of multiple simultaneous or related defects or other events.

In embodiments, the digital twin behavior model may be updated and improved using results from actual experiments and real-world events. The use of these digital twin mathematical models and simulations avoids real-world experiments, which can be costly and time-consuming. Instead, the acquired knowledge about the entity's behavior and computational capabilities is used to diagnose and solve real-world problems in an inexpensive and/or time-efficient manner. As such, digital twin simulation system 8116 can facilitate understanding the behavior of a system without actually testing the system in the real world. For example, during the design of a tractor to determine which type of wheel configuration would best improve traction, a digital twin model simulation of the tractor could be used to estimate the effect of different wheel configurations on towing ability. Useful insights into different decisions in the design can be gained without actually building the tractor. Additionally, digital twin simulations can support experiments that occur entirely in software or in a human-in-the-loop environment where the digital twin represents the system or generates the data needed to satisfy the experiment objectives. Additionally, digital twin simulations can be used to train people using appropriate virtual environments in aspects that would otherwise be difficult or expensive to create.

In embodiments, a simulation environment may be constructed using a model configured to predict a set of future states. These models may include deep learning, regression models, quantum prediction engines, inference engines, pattern recognition engines, and many other forms of modeling engines that use past outcomes, current state information, and other inputs to build future state predictions. You can. In some embodiments, considerations in creating the functionality of the digital twin model include the ability to also show the response of perspective-based digital twin structural elements (e.g., defining the deformation of a vehicle's axle in response to different magnitude loads). am. For example, the resulting digital twin representation may then be presented to the user in a virtual reality or augmented reality environment where a particular viewpoint is depicted in digital twin form.

In embodiments, a digital twin may operate in concert with an adaptive edge computing system and/or a set of adaptive edge computing systems that provide coordinated edge computation, as described herein, including a wide range of systems: For example, classification systems (e.g., image classification systems, object type recognition systems, etc.), video processing systems (e.g., video compression systems), signal processing systems (e.g., analog-to-digital conversion systems, digital-to-analog conversion systems, RF filtering systems, Analog signal processing systems, multiplexing systems, statistical signal processing systems, signal filtering systems, natural language processing systems, sound processing systems, ultrasonic processing systems, etc.), data processing systems (e.g., data filtering systems, data integration systems, data extraction systems, data Loading system, data transformation system, point cloud processing system, data normalization system, data cleansing system, data deduplication system, graph-based data storage system, object-oriented data storage system, etc.), prediction system (e.g., motion prediction system, output prediction) systems, activity prediction systems, fault prediction systems, failure prediction systems, accident prediction systems, event prediction systems, event prediction systems, etc.), configuration systems (e.g., protocol selection systems, storage configuration systems, peer-to-peer network configuration systems, power management systems, self-organizing systems, self-healing systems, handshake negotiation systems, etc.), artificial intelligence systems (e.g., clustering systems, variational systems, machine learning systems, expert systems, rule-based systems, deep learning systems, etc.), System management and control systems (e.g., autonomous control systems, robot control systems, RF spectrum management systems, network resource management systems, storage management systems, data management systems, etc.), robotic process automation systems, analysis and modeling systems (e.g., data visualization) systems, clustering systems, similarity analysis systems, random forest systems, physical modeling systems, interaction modeling systems, simulation systems, etc.), entity discovery systems, security systems (e.g., cybersecurity systems, biometric systems, intrusion detection systems, firewall systems) etc.), rule engine system, workflow automation system, opportunity discovery system, testing and diagnostic system, software image propagation system, virtualization system, digital twin system, IoT monitoring system, routing system, switching system, indoor location system, geolocation system. Includes etc.

In an embodiment, digital twin notification system 8118 provides notifications to users via the enterprise digital twin associated with each user. In some embodiments, digital twin notifications are an important part of the overall interaction. The digital twin notification system 8118 can provide digital twin notifications within the context of a digital twin setting, such that a perspective view of the notification allows for understanding how the notification fits into the general digital twin represented ontology, taxonomy, topology, etc. It is specifically set to do so.

As discussed, the digital twin model is based on a combination of data and its relationship to the digital twin environment and/or process. As such, different digital twins may share the same data, and different digital twin perspectives may be the result of a digital twin data model or set of metadata built on top of the data environment. In embodiments, a digital twin data model provides details of the information to be stored and is used to build a layered system that can represent low-level information in a form appropriate to the digital twin perspective for which the final computer software code will be used. . One aspect of a digital twin model is that a single digital can be shared across multiple viewpoints, and each viewpoint viewer can then interact with the same underlying digital twin model. In this way, multiple perspectives are similar to transformations that allow each type of user to interact in a manner appropriate for their skill set or knowledge level.

70 illustrates an example of how a digital twin data model and digital twin are created, executed, and served to a requesting digital twin application, where the digital twin data model is drawn from existing systems and digital twin structures to achieve a digital twin representation. Defines the physical implementation of the underlying data stream. In an embodiment, digital twin data model 81B00 defines how traditional data streams are linked together with a digital twin structure to achieve a digital twin representation. In embodiments, a digital twin is a combination of process/structure and system data streams. In other words, process and structure definitions are either real-world “things” (e.g. factories, robots, cargo containers, ships, roads, etc.) or logical “things” (e.g. organizational charts, hiring processes, marketing campaigns, tax reporting workflows, etc.), while system data stream definitions define how real-world data can be collected into digital twin representations of real-world and/or logical “things.” Therefore, constructing a digital twin includes structural construction and collection and data construction and collection.

During structural construction and collection, digital twin system 8004 receives structural aspects of the digital twin. In embodiments, structural aspects may include process definitions, layout definitions, and/or spatial definitions. In embodiments, a process definition defines a logical process that can be mapped to a high-level format that forms the basis of what a digital twin viewer can interact with. Examples of processes may include workflows, hiring processes, manufacturing processes, logistics processes, inventory processes, product management processes, software processes, etc. In embodiments, the spatial definition defines the geospatial organization of an object or environment. In embodiments, the spatial definition may be a 2D or 3D representation of an object or environment. Spatial definitions of objects or environments may be provided as CAD files, LIDAR scans, 2D or 3D images, etc., including logical relationships, organizational hierarchies, physical relationships, schematic relationships, and/or interconnections between objects and/or environments. . In embodiments, a layout definition defines relationships between objects and other objects and/or the environment. In embodiments, the layout definition may further define how objects move relative to other objects and/or the environment. Examples of layouts may include electrical wiring diagrams, plumbing outlines, assembly line diagrams, circuit diagrams, hierarchical relationships, network layouts, network outlines, organizational charts, etc. In embodiments, a layout definition may include a set of characteristics of an object or environment. Examples of properties of an object include physical properties such as the material of the object, the weight of the object, the density of the object, the conductivity of the object, the resistance of the object, the maximum velocity of the object, the maximum acceleration of the object, the possible movements of the object, the reactivity of the object, etc. can do. Examples of characteristics of the environment may include materials of floors, walls, roofs, etc., the coefficient of friction of the floor, confined areas within the environment, paths within the environment, and/or other suitable characteristics. In some embodiments, users may upload layout definitions, process definitions, and/or space definitions to digital twin system 8004. Additionally or alternatively, digital twin system 8004 may provide a graphical user interface that allows users to define layout definitions, process definitions, and/or space definitions. In some embodiments, users can import digital twins from third-party sources. For example, the producer of a particular object may also provide a digital twin of the object, which may then be imported into the digital twin system 8004.

During system data configuration and collection, users define data sources that provide data to hydrate or populate the digital twin, and configure data buses to receive data from various data sources. As discussed, data sources may include sensor systems, ERP, CRM, financial systems, inventory management systems, invoicing systems, third party systems (e.g., weather services, news services, government databases, etc.), and other suitable systems. It can be received from various systems including. In embodiments, a user may identify a data source and provide any information required to enable a data bus to receive data from the data source and between data derived from the data source and the digital twin element. The correlation can be further defined. A data bus may refer to a middleware layer that provides data wiring and data infrastructure for moving data from one system to another. The data bus may be configured to handle real-time data, near real-time data, aggregated data, and/or stored data, or any combination thereof. The data bus can provide data directly to the digital twin and/or store data in a data warehouse that hydrates the digital twin. In embodiments, a user may enable data acquisition from a data source by providing an API interface or key and/or webhook URL to the digital twin system 8004 (e.g., via a GUI). In embodiments, digital twin system 8004 may configure a data bus to access and/or receive data from data sources. In some of these embodiments, digital twin system 8004 can generate a webhook URL for a particular digital twin or set of digital twins, allowing data sources to push real-time or near-real-time data to a data bus. A URL can be provided to the data source. Additionally or alternatively, digital twin system 8004 can obtain an API interface or key from a data source so that the data bus can request data from the data source using the API interface or key.

In embodiments, digital twin system 8004 may generate foreign keys that associate different types of data with structural elements of the digital twin. In this way, when a digital twin is displayed, foreign keys link specific data types to various structural or logical or schematic elements so that real-world data collected from various data sources is linked to the corresponding state of the digital twin. For example, sensor data received from a subset of sensors in a sensor system monitoring a particular mechanical component in a real-world environment may be associated with the mechanical component's digital twin, such that the sensor data may be displayed in the mechanical component's digital twin. . In embodiments, users may provide input to digital twin system 8004 during the configuration phase to associate specific data types to various elements of the digital twin. Data types associated with a digital twin may include raw data, processed data, analyzed data, derived data, etc. As long as certain data streams are processed before being served to the digital twin (for example, sensor data averaged over a period of time or an alert condition displayed when sales data falls below a threshold), users can ensure that the data is served to the digital twin. You can define associated display highlights or actions performed on data before it is displayed. In this scenario, processed data can be associated with each digital twin component within a foreign key.

Once a data bus is configured for a particular digital twin and the structural, logical, or schematic elements of the digital twin (e.g., layout definition, process definition, and spatial definition) are defined, the digital twin system 8004 A digital simulation may be performed and/or the digital twin may be served to a digital twin-enabled application based on the digital twin's structural elements, connected system data sources, and the digital twin's external keys. In embodiments, the digital twin may be a role-based digital twin, whereby views into the digital twin are served to users occupying specific roles within the organization. In this way, each user can interact with each role-based digital twin and gain an appropriate perspective based on their respective needs for the organization. In another embodiment, multiple users can interact with a shared role-enabled digital twin and obtain an appropriate perspective based on their respective needs regarding the organization of the single digital twin. In embodiments, a role-based digital twin may allow users to provide feedback to the source system, allowing control of the source system environment, such as corrective actions taken against the source system. In some embodiments, multiple users can make behavioral changes to a shared role-based digital twin, and each user can view these changes in a manner appropriate to their role. Additionally, when behavioral changes involve multiple users, the digital twin can enable role-based workflow management of the displayed environment (for example, the CEO may spend money to change a machine as requested by the CTO). can be approved).

In embodiments, digital twin system 8004 may receive a request to run a digital twin simulation on a digital twin. A request to perform a digital twin simulation may be received from a digital twin application and/or from an internal process. In embodiments, digital twin simulation allows for the construction of interaction models based on the process, layout, and/or spatial representation of the digital twin. Digital twin simulations can provide degrees of freedom to allow different processes to change in response to dynamic data input. For example, a digital twin simulation could be run to display how a bearing may move on a compressor when the compressor is operated under different operating conditions, or how water flows through a system in a pipe model at different temperatures or with different accumulations within the pipe. You can. In embodiments, digital twin system 8004 may output results of a simulation that may indicate, for example, the impact of simulation parameters on certain aspects of the digital twin.

In embodiments, a digital twin application may request and display a digital twin to a user, which may be a new twin for that user or a role-specific access with a role-specific view of an existing or shared digital twin. there is. Digital twin applications can be provided on mobile applications, virtual reality applications, PCs, etc. In an embodiment, a digital twin application provides a request for a specific digital twin to digital twin system 8004, where the request may include the user's user identifier and/or the user's role. In embodiments, digital twin system 8004 may include or interface with a digital twin application coordinator that receives requests from digital twin applications for digital twins. In embodiments, the digital twin application controller maintains and utilizes a set of business rules for the specific digital twin required by the digital twin application. In some of these embodiments, the set of role-based rules is a set of role-based rules that control what roles a user has within an organization and what states the user can access given the user's permission. In this embodiment, the digital twin application controller may determine whether to grant an instance of digital twin application access to a particular user based on business rules and the user's role. In embodiments, digital twin system 8004 may include an application services layer that allows multiple users to connect to the backend of a digital twin application coordinator directly or through a shared digital twin. In embodiments, these connections may include web services, public and subscription information buses, simple object access protocols, and/or other suitable application interfaces. The application service layer can return the requested digital twin as a request instance of the digital twin application, which in turn displays the digital twin to the user. The user can then interact with the digital twin through the application to view different states of the digital twin, request simulations, or interact with other users in the same or different roles in the digital twin environment, etc.

In an example implementation of the framework discussed in FIG. 70, digital twin system 8004 can be configured to create an enterprise digital twin with respect to the value chain. For example, a company that produces goods internationally (or at multiple facilities) may have supplier twins that represent the company's supply chain, factory twins at various production facilities, product twins that represent products made by the company, and the company's distribution chain. A set of digital twins can be constructed, such as a distribution twin representing a distribution twin, and other suitable twins. In doing so, companies can define the structural elements of each individual digital twin, as well as arbitrary system data that corresponds to the structural elements of the digital twin. For example, in creating a production facility twin, a company can shape the layout and spatial definition of the facility and any processes performed in the facility. Companies can also define data sources that correspond to value chain entities such as sensor systems, smart manufacturing equipment, inventory systems, logistics systems, etc. that provide data related to facilities. Companies can associate data sources with elements of a production facility and/or processes that occur in the facility. Similarly, enterprises can define the structural, process, and layout definitions of their supply chains and distribution chains, and connect relevant data sources such as supplier databases and logistics platforms to create respective distribution chains and supply chain twins. . Companies can further relate these digital twins to have a view of the value chain. In embodiments, digital twin system 8004 may perform a simulation of an enterprise's value chain incorporating real-time data obtained from the enterprise's various value chain entities. In some of these embodiments, digital twin system 8004 may determine when to order specific parts to manufacture a specific product, when to schedule maintenance on a machine, and/or, given the predicted demand for the manufactured product. Whether to replace a machine (for example, when a digital simulation of a digital twin indicates there may be the least demand for a particular product or when it will have the least impact on a company's income statement), and which day to ship the item. Decisions, such as time to use, can be recommended to users interacting with the enterprise digital twin. The foregoing example is a non-limiting example of how a digital twin can collect system data and perform simulations to add one or more objectives.

Figure 71 shows examples of different types of enterprise digital twins, including executive digital twins, in relation to the data layer, processing layer, and application layer of the enterprise digital twin framework. In embodiments, the executive digital twin includes CEO digital twin (8302), CFO digital twin (8304), COO digital twin (8306), CMO digital twin (8308), CTO digital twin (8310), CIO digital twin (8312), It may include, but is not limited to, GC digital twin (8314), HR digital twin (8316), etc. Additionally, enterprise digital twins that may be associated with the executive suite may include cohort digital twins (8320), agility digital twins (8322), CRM digital twins (8324), etc. The discussion of different types of digital twins is provided as an example and is not intended to limit the scope of the present disclosure. It is understood that in some embodiments, users may change the configuration of various executive digital twins based on the business needs of the enterprise, the reporting structure of the enterprise, and the roles and responsibilities of various executives within the enterprise.

In embodiments, executive digital twins and additional enterprise digital twins are created using various types of data collected from different data sources. As discussed, data can be real-time data (8330), historical data (8332), analytical data (8334), simulated/modeled data (8336), CRM data (8338), organizational charts, and/or organizational digital twins (8340). It may include organizational data, enterprise data lake 8342, and market data 8344. In an embodiment, real-time data 8330 may be transmitted directly from each sensor and/or reader (e.g., RFID, NFC, and Bluetooth reader), beacon, gateways, repeater, mesh network node, WIFI system, access point, It may include sensor data collected from one or more IoT sensor systems, which may be collected by various data collection devices associated with the enterprise, including routers, switches, gateways, local area network nodes, edge devices, etc. Real-time data 8330 may include additional or alternative types of data collected in real-time, such as real-time sales data, real-time cost data, project management data indicating the status of the current project, etc. Historical data may be any data collected by and/or on behalf of a business in the past. This includes sensor data collected from the enterprise's sensor systems, sales data, cost data, maintenance data, purchasing data, employee employment data, employee on-boarding data, employee retention data, legal-related data indicating legal proceedings, patent applications and It may include patent application data showing granted patents, project management data showing past progress on past and current projects, product data showing products on the market, etc. Analytical data 8334 may be data derived from performing one or more analytical processes on data collected by and/or on behalf of an enterprise. Simulated/modeled data 8336 may be any data derived from a simulation and/or behavior modeling process performed on one or more digital twins. CRM data 8336 may include data obtained from the company's CRM. The organizational digital twin 8340 may be a corporate digital twin. Enterprise data lake 8342 may be a data lake that includes data collected from any number of sources. In embodiments, market data 8342 may include data collected from disparate data sources about or related to competitors and other cohorts within the market and supply chain. Market data 8342 may be collected from many different sources and may be structured or unstructured. In embodiments, market data 8342 may include an element of uncertainty that may be displayed in a digital twin relying on such market data 8342, for example, by showing error bars, probability cones, random walk paths, etc. You can. It is understood that the different types of data highlighted above may overlap. For example: historical data can be obtained from CRM data; Enterprise data lake 8342 may include real-time data 8330, historical data 8332, analytical data 8332, simulated/modeled data 8336, and/or CRM data 8336; Analytics data 8334 may be based on historical data 8332, real-time data 8332, CRM data 8336, and/or market data 8342. Additional or alternative types of data may be used to populate the enterprise digital twin.

In embodiments, data structuring system 8106 may structure various data collected by and/or on behalf of an enterprise. In an embodiment, digital twin creation system 8108 creates an enterprise digital twin. As discussed, digital twin creation system 8108 may receive a request for a particular type of digital twin (e.g., CEO digital twin 8302 or CTO digital twin 8310), and may Based on the composition of the digital twin, you can determine the type of data needed to populate the digital twin. In embodiments, digital twin creation system 8108 may then create a requested digital twin based on various types of data (which may include structured data structured by data structuring system 8106). In some embodiments, digital twin creation system 8108 can output the generated digital twin to a client application 8052, and client application 8052 can then display the requested digital twin.

In an embodiment, CEO digital twin 8302 is a digital twin configured for an enterprise's CEO or similar top-level decision maker. CEO Digital Twin 8302 is an enterprise's digital twin that includes real-time and historical representations of key assets, processes, divisions, performance metrics, conditions of the enterprise's different business units, and any other mission-critical information types. May include high-level views of different status and/or operational data. In embodiments, CEO digital twin 8302 may utilize simulations, predictions, statistical summaries, analytics, machine learning, and/or other AI-based decision-support and input (e.g., financial data, competitor data, product data, etc.). can work in conjunction with EMP (8000) to provide learning-type processing of In embodiments, CEO digital twin 8302 may be responsible for managing people, delegating tasks, making decisions or tasks, coordinating with the board of directors and/or strategic partners, managing risk, managing policies, overseeing budgets, allocating resources, investing, and other tasks. May provide functionality including, but not limited to, enforcement-related resources.

In embodiments, the types of data that may populate CEO digital twin 8302 may include, but are not limited to: macroeconomic data, microeconomic analysis data, forecast data, demand planning data, and employment and payroll data. , analytical results from AI and/or machine learning modeling (e.g., financial projections), forecast data, recommendation data, securities-related financial data (e.g., revenue, profitability), industry analyst data (e.g., Gartner quadrants) ), strategic competitive data (e.g., news and events regarding industry trends and competitors), business performance metrics by business units that may be relevant to evaluating the performance of a business unit (e.g., P&L, head count, factory health, supply chain metrics, sales metrics, R&D metrics, marketing metrics, etc.), board package data, or some other type of data related to the operations of the CEO and/or executive department. In embodiments, digital twin system 8004 may collect securities-related financial data, for example, from a company's accounting software (e.g., via an API), publicly disclosed financial statements, third-party reporting, tax returns, etc. It can be obtained. In embodiments, digital twin system 8004 may obtain strategic competitive data from public news sources, publicly disclosed financial reports, etc. In embodiments, macroeconomic data may be analytically derived from various financial and operational data collected by EMP 8000. In embodiments, business performance metrics may be analytically derived by the artificial intelligence service system 8010 based at least in part on real-time operational data and/or provided from other users and/or their respective executive digital twins. CEO Digital Twin 8302 defines real-time operational data parameters of interest and monitors real-time operational data for conformance and alignment with the organization's stated business objects, board requirements, industry best practices, regulations, or some other criteria; It can be used to collect, analyze, and interpret.

In embodiments, CEO digital twin 8302 may include a high-level view of the different states of the enterprise, including real-time and historical representations of key assets, conditions of the enterprise's different business units, and any mission-critical information. You can. The CEO digital twin 8302 may initially display various states at a lower level of granularity. In embodiments, a user viewing the CEO digital twin 8302 can select states to drill down to and view the selected states at a higher level of granularity. For example, CEO digital twin 8302 initially displays a subset of the various states of the enterprise at a lower level of granularity, including the state of the finance department (e.g., a visual indicator of the enterprise's overall financial health score). can do. In response to the selection, CEO digital twin 8302 may include real-time, historical, aggregated, comparative, and/or projected financial information (e.g., real-time, historical, simulated, and/or projected revenue, liabilities, etc.) However, we may provide, but are not limited to, data, analysis, summaries, and/or reports. In this way, the CEO digital twin 8302 may initially provide a user (e.g., a CEO) with a view to displaying various different aspects of the enterprise (e.g., different “health” levels of each business unit or portion of the enterprise). can present views of different indicators), but allow the user to choose which aspects need more attention. In response to this selection, CEO digital twin 8302 may request a more granular view of the selected state(s) from EMP 8000, which may return the requested state at a more granular level.

In embodiments, CEO digital twin 8302 may include executive-level digital twins of executive departments (e.g., C-suite, directors, board members, etc.), where users may be members of the board of directors involved in the oversight of the organization's governance. It can be used to identify, assign, direct, supervise and review executive department personnel and third party personnel, departments, organizations, etc., associated with the activities of the organization's officers, including In embodiments, an executive-level digital twin may include definitions of the various roles, employees, and departments working under the CEO, a reporting structure for each individual within the business unit, and various names and/or names of the individuals filling each role. Or it can be filled with another identifier. In embodiments, CEO digital twin 8302 defines/redefines workforce groupings, assigns performance criteria and metrics to business units, roles, and/or individuals, and/or tasks the business units through the executive-level digital twin. , roles, and/or individuals, etc. may include a graphical user interface that provides the user with the ability to assign/delegate. In embodiments, an executive-level digital twin may provide real-time operational data for an organization to continuously evaluate the performance of workforce groupings against stored performance criteria.

In embodiments, CEO digital twin 8302 may be configured to interface with collaboration suite 8006 to specify and provide a set of collaboration tools that executive departments and associated parties can utilize. Collaboration tools include, as described herein, videoconferencing tools, “in-twin” collaboration tools (e.g., where collaboration occurs to some extent within a common interface where digital twin entities are observed and collaboration activities occur) and/or where the components of the EMP used to configure, operate, or support the digital twin also govern collaboration around the digital twin entities and workflows), whiteboard tools, agile development environment tools (e.g., in a Slack™ environment) features), presentation tools, word processing tools, spreadsheet tools, etc. Collaboration and communication rules may be constructed based at least in part on using an AI reporting tool, as described herein. Collaboration tools facilitate collaborative communication (e.g., live meetings where participants are simultaneously presented with meeting-related views or workflows of digital twin entities), asynchronous collaboration (e.g., actions, comments, etc. on digital twin entities) presented to different users interacting with it), version control features, etc.

In embodiments, CEO digital twin 8302 may be responsible for overall strategic objectives, policy implementation, product rollouts, board interactions, investments or acquisitions, investor relations, public relations and media handling, budgeting, or some other type of executive initiative. , but may be configured to provide research, tracking, and reporting on executive branch initiatives, including but not limited to: The CEO Digital Twin (8302) can interact with and share data and reporting with other Executive Digital Twins, including but not limited to CFO Digital Twin, COO Digital Twin, etc. In embodiments, CEO digital twin 8302 or an enforcement agent integrated with or within it (e.g., trained to perform expert enforcement actions as described elsewhere in this application) may use intelligent services (e.g. , data analytics, machine learning, and A.I. processes) to analyze financial reports, projections, simulations, budgets, and related summaries, including key departments, personnel, and resources listed or subordinate to projects, initiatives, budget line items, etc. , may identify a third party or other person, and he may therefore be interested in such data. Such material regarding a given party may be abstracted and summarized for presentation, formatted and presented automatically, or at the direction of the CEO or other user, to the party attributable to the cost and/or subject matter of the material. For example, CEO digital twin 8302 may assemble materials for the purpose of developing presentations, speaking points, press releases, or some other materials for the CEO or other executive personnel to use for public presentations. In an example, a CEO expected to give a conference presentation regarding the introduction of a new company product may use the CEO digital twin 8302 to obtain product data (e.g., units produced, units shipped), financial data (e.g., (e.g., products sold, reserved products), graphical presentation information (e.g., product photos, maps of product distribution, graphs of projected sales), forecast data (e.g., projected market growth), or portions thereof. Specifies and organizes the identification, collection and assembly of operational data relevant to the upcoming presentation, such as other types of data, presentation slides, white paper templates and remarks that may form the basis of the presentation or be distributed in conjunction with the presentation and/or its marketing; You can assemble that information into a presentation format, such as a talking point, press release, or some other summary format.

In embodiments, CEO digital twin 8302 may be configured to track and report stakeholder communications (e.g., reports, board requests, investor requests) related to executive departments. The CEO Digital Twin (8302) can present, store, analyze, coordinate and/or report on enforcement activities related to parties with whom the enforcement department is contracting, collaborating, reporting, etc., such as key personnel, external contractors, media, board of directors, etc. there is.

In embodiments, CEO digital twin 8302 may be configured to simulate one or more aspects of an enterprise. Such simulations can assist users (e.g., CEOs) in making executive level decisions. For example, a simulation of a proposed executive initiative may include, for example, the initiative (e.g., introduction of a new product), various financial parameters (e.g., potential investment levels), and targeting parameters, as described herein. It can be tested using modeling, machine learning, and/or AI techniques, by simulating temporal effects on (e.g., geographic, demographic, etc.), and/or other suitable enforcement parameters. In embodiments, digital twin simulation system 8116 may receive a request to perform an executive simulation requested by CEO digital twin 8302, where the request indicates one or more parameters to be changed in one or more enterprise digital twins. . In response, the digital twin simulation system 8116 may return simulation results to the CEO digital twin 8302, which in turn outputs the results to the user through the client device display. In this way, the user can be provided with a variety of results corresponding to different parameter configurations. For example, a user can request a set of simulations to be run to test different supply chain strategies to see how different strategies affect a manufacturing facility's throughput and the overall impact on the company's profits and losses. The digital twin simulation system 8116 can perform simulations by varying different supply chain strategies and output throughput and P&L projections for each supply chain strategy. In some embodiments, a user may select a set of parameters based on various results and repeat the simulation based on at least the changed previous results. From the previous example, the user may decide to select a supply chain strategy that maximizes P&L projections but does not adversely affect the manufacturing facility's throughput. In some embodiments, an enforcement agent may be trained to recommend and/or select parameter sets based on respective results associated with each respective parameter set.

In embodiments, CEO digital twin 8302 may be configured to store, aggregate, merge, analyze, prepare, report, and distribute data related to execution strategies, execution plans, execution activities, and/or execution initiatives. For example, CEO digital twin 8302 can provide financial data, summaries, and reports and analyzes related to previous executive activities (e.g., previous quarterly financial performance, previous investments, previous strategic partners, joint developments, etc.). may be associated with a plurality of databases or other repositories of financial data, summaries and reporting and analysis, each of which may be further associated with financial and performance metrics related to the campaign and may also be associated with CEO digital twin 8302. It is accessible by

In embodiments, CEO digital twin 8302 stores, aggregates, merges, analyzes, prepares, reports, and distributes material related to financial reporting, ratings, rankings, financial trend data, earnings data, or other data related to executive responsibilities. It can be configured to do so. CEO digital twin 8302 can connect to, interact with, and associate with external data sources, and upload and download external data sources, including internal data of the EMP, as described herein. , aggregate, and analyze this data. Data analytics, machine learning, AI processing, and other analytics may be coordinated between CEO digital twin 8302 and the analytics team based at least in part on using artificial intelligence service system 8010. These collaborations and interactions seed enforcement-related data elements and domains into the enterprise data repository 8012 for use in modeling, machine learning, and AI processing to identify optimal business strategies, or some other enforcement-related metric or aspect. In addition, it may also include assisting in the identification of optimal data measurement parameters on which to base judgments on the success of an executive initiative. Examples of data sources 8020 that can be connected to, associated with, and/or accessed from CEO digital twin 8302 include facilities (e.g., manufacturing facilities, shipping and logistics facilities, transportation facilities, agricultural facilities) , resource extraction facilities, computing facilities, etc.) and/or other physical entities of the enterprise, including a sensor system 8022, a sales database 8024 updated with sales figures in real time, and a CRM system 8026. , content marketing platforms (8028), financial databases (8030), surveys (8032), organizational charts (8034), workflow management systems (8036), third-party data sources (8038), and customer databases that store customer data (8038). 8040), and/or third-party data sources 8038 that store third-party data, edge devices 8042 that report data related to physical assets (e.g., smart machines/manufacturing equipment, sensor kits, enterprise may include, but are not limited to, autonomous vehicles, wearable devices, etc.), an enterprise resource management system (8044), an HR system (8046), a content management system (8016), etc. In embodiments, digital twin system 8004 abstracts the different views (or states) within the digital twin to appropriate granularity. For example, digital twin system 8004 may have access to real-time sensor data streams as well as access to all sensor data collected on behalf of the enterprise. Typically, this data is too granular for an executive like the CEO, and sensor data readings are often of little importance to the CEO unless they are associated with mission-critical conditions or actions. However, in this example, if sensor readings from a particular physical asset (e.g., a critical piece of manufacturing equipment) indicate a potentially critical situation (e.g., a fault condition, hazardous condition, etc.), then analysis can become very important to CEOs. Accordingly, the digital twin system 8004 may include health indicators of the physical assets in the CEO digital twin when building an appropriate view of the CEO. In this way, CEOs can drill down into health indicators of physical assets to view potentially critical situations (e.g., analysis of sensor data and machines used to identify situations) with greater granularity.

In an embodiment, the CEO digital twin 8302 organizes the organization based at least in part on real-time operational data and usage of a monitoring agent of the client application 8052, as described herein, associated with the CEO digital twin 8302. It can be configured to monitor performance. The monitoring agent may report on these activities to EMP 8000 for presentation in a user interface associated with CEO digital twin 8302. In response, EMP 8000 trains an enforcement agent (which may include one or more machine learned models) to handle and process these notifications the next time they arrive, and determine if these notifications are of an urgent nature. When, for example, the announcement of an acquisition by a competitor, a report flagging an underperforming business unit, a high-profile press article, a dramatic change in the stock market (for the CEO's company, a member of the cohort, or the market as a whole), or a rating by an industry analyst. It can alert the Sangsin and/or the CEO of a downgrade, an external event that has the potential to disrupt operations (such as a natural disaster or epidemic), or some other important event. In embodiments, CEO digital twin 8302 may generate performance alerts based on real-time operational data, performance trends, etc. This can allow CEOs to optimize initiatives in real time without having to manually request such real-time data; CEO digital twin 8302 may automatically present such information and relevant/necessary alerts as configured by the organization, the CEO, or some other interested party.

In embodiments, CEO digital twin 8302 may be configured to report on executive department performance, executive department personnel, enforcement activities, enforcement content, enforcement platforms, enforcement partners, or some other aspect of management within the CEO's responsibilities. there is. Reports may be to the CEO, executive departments, other executives of the organization (e.g., COO), or external third parties (e.g., partners, press releases, etc.). As described herein, reporting may include stakeholder summaries, minutes of meetings, presentations, sales data, customer data, financial performance metrics, workforce metrics, data regarding resource use, industry summaries (e.g., may include summaries of merger and acquisition activity), or some other type of reporting data. The report and its contents may be shared by the CEO digital twin 8302 with other executive digital twins. The reporting capabilities of CEO Digital Twin 8302 can also be used to fill new or pre-set reporting formats, etc. Templates of common reporting formats can be stored and associated with the CEO digital twin (8302) to automate the presentation of data and analysis according to predefined formats, styles and system requirements. In embodiments, an enforcement agent trained by a user may be trained to surface the most important reports to the user. For example, if a user (e.g., a CEO) consistently views and tracks sales data reporting but routinely skips reporting related to manufacturing KPIs, an enforcement agent can automatically surface sales data reporting to the user and KPIs can be automatically delegated to another executive digital twin (for example, the COO digital twin).

In embodiments, CEO digital twin 8302 may be configured to monitor, store, aggregate, merge, analyze, prepare, report, and distribute data related to the CEO's organization or competitors of the named entity of interest. In embodiments, such data may be retrieved from sources including, but not limited to, information regarding investments and/or acquisitions, press releases, SEC or other financial reports, or some other publicly available data; Data may be collected by EMP 8000 through data aggregation, spidering, web-scraping, or other techniques. For example, a user wishing to monitor a specific competitor could request that CEO digital twin 8302 provide data related to that specific competitor. In response, EMP 8000 may identify a set of data sources that are publicly available or accessible to the CEO's enterprise (e.g., internal data sources, authorized third party data, etc.). EMP 8000 may construct a cohort digital twin 8320 based on the type of data/analysis/service of the user request and the set of identified data sources. EMP 8000 may then serve the CEO digital twin 8302 with a cohort digital twin 8320 associated with the requested party (e.g., a competitor).

In embodiments, CEO digital twin 8302 may be configured to monitor, store, aggregate, merge, analyze, prepare, report, and distribute material related to regulatory activities, such as government regulations, industry best practices, or some other requirements or standards. there is. For example, the CEO digital twin (8302) can communicate with other corporate digital twins, such as the general counsel digital twin (8314), which allows the legal team to keep the CEO informed when new regulations or regulatory changes occur.

In an embodiment, a client application 8052 executing the CEO digital twin 8302 may be configured with an executive agent 8364 that is trained on the CEO's actions (which may indicate behaviors and/or preferences). In embodiments, executive agent 8364 may record action-related characteristics (e.g., context related to the user's action) to expert agent system 8008. For example, whenever a user delegates a task to a subordinate (which is an action), the executive agent 8364 may determine characteristics surrounding the delegation of the task (e.g., the event that caused the user to delegate the task, the type of task delegated, The role to which the task is delegated, the commands provided by the user along with the delegation, etc.) can be recorded. Executing agent 8364 may report actions and characteristics to expert agent system 8008, which may then execute in such a way that executing agent 8364 may delegate or recommend delegation of the task in the future. The agent 8364 can be trained. Once trained, executive agent 8364 can automatically perform actions and/or recommend actions to the user. Additionally, in embodiments, executive agent 8364 may record results associated with performed/recommended actions, thereby creating a feedback loop with expert agent system 8008.

References to the features and functionality of EMP and digital twins in this example of CEO digital twin 8302 apply to other digital twins and their respective projects and workflows, unless the context indicates otherwise. It must be understood that

In embodiments, Chief Financial Officer (CFO) digital twin 8304 may be a digital twin configured for an enterprise's CFO or similar executive responsible for overseeing financial-related tasks for the enterprise. CFO digital twin 8304 includes real-time, historical, aggregated, comparative, and/or projected financial information (e.g., real-time, historical, simulated, and/or projected sales figures, expenses, revenue, liabilities, etc.) However, we may provide, but are not limited to, data, analysis, summaries, and/or reports. In embodiments, the CFO digital twin operates in conjunction with EMP 8000 to perform simulations, forecasts, statistical summaries, analytics, machine learning, and/or other AI and inputs (e.g., accounting data, sales data, sensor data, etc.). ) can provide decision support based on learning-type processing.

In embodiments, CFO digital twin 8304 may be responsible for managing finance staff, partners, and external consultants and contractors (e.g., accounting firms, auditors, etc.), overseeing budgets, procurement, expenditures, receivables, and other finance-related resources; Compliance, supervision of sales and financial performance of sales personnel and departments, administration of contracts, administration of internal policies (e.g., policies related to expenditures and reporting), tax laws, and financial privacy laws (e.g., credit agency data and may provide features and functionality including, but not limited to, reporting, compliance, and regulatory analytics (related to this).

In embodiments, the types of data that can populate a CFO digital twin include: financial performance metrics, revenue, cash, and balance sheet data by business unit, by product, by geography, by factory, by store location(s), by asset class; Cash flow, profitability, resource utilization, audit data, general ledger data, asset performance data, securities and commodities data, insurance and risk management data, asset aging and depreciation data, asset allocation data, macroeconomic data, microeconomics. Analytics data, tax data, pricing data, competitive product and pricing data, forecast data, demand planning data, employment and salary data, analytical results from AI and/or machine learning modeling (e.g. financial projections), forecast data, recommendations. may include, but is not limited to, data, or some other type of data related to the operations of the CFO and/or finance department. In embodiments, when used in this application, “datum,” “data,” “data set,” “datastore,” “data warehouse,” and/or “database” means to summarize, input, or use statistical or scientific notation. It can refer to information stored in a numerical or statistical format, including output, information stored in natural language format (e.g. text extracts from reports, press releases, legislation, etc.), and information stored in graphical format. (e.g., financial performance graphs), information stored in audio and/or audio-visual format (e.g., recorded information from conference calls, including natural language transcription summaries of the information in audio and/or audio-visual format). audio or video from a presentation), or some other type of information.

In embodiments, CFO digital twin 8304 may display a finance department twin of a finance department, allowing users to identify third-party partners, such as accounting firms, tax attorneys, etc., and other external contractors involved in the organization's financial efforts. , can be used to identify, assign, direct, supervise, and review finance department personnel and third-party personnel associated with an organization's financial activities. Examples of such organizational personnel include, but are not limited to, finance department personnel, sales analysts, statisticians, data scientists, executive personnel, human resources staff, board members, advisors, or some other type of organizational personnel associated with the functions of the finance department. . Examples of third-party personnel in the finance department include, but are not limited to, attorneys, accountants, management consultants, social media platform personnel, finance partners, consultants, contractors, finance firm personnel, auditors, or some other type of third-party personnel. .

In embodiments, CFO digital twin 8304 may include definitions of the various roles/employees working under the CFO, reporting structures, and associated permissions for each individual within the business unit, and may include a variety of individuals filling each role. It may be filled with a name and/or other identifier. In embodiments, a user (e.g., a corporate CFO) may use CFO digital twin 8304 to coordinate the reporting structure within the finance department and/or grant permissions to one or more individuals within the department.

In embodiments, CFO digital twin 8304 may be configured to interface with collaboration suite 8006 to specify and provide a set of collaboration tools that the finance department and associated parties can utilize. Collaboration tools may include video conferencing tools, “in-twin” collaboration tools, whiteboard tools, presentation tools, word processing tools, spreadsheet tools, etc., as described herein. Collaboration and communication rules may be constructed based at least in part on using an AI reporting tool, as described herein.

In embodiments, CFO digital twin 8304 may be used by a finance department, including, but not limited to, an overall department budget, a budget for a single or group of financial initiatives, audits, third-party vendor activities, or some other type of cost or budget. It can be configured to research, create, track and report on initiatives. In embodiments, CFO digital twin 8304 may include, but is not limited to, a digital twin associated with accounts payable, an executive employee, such as a CEO (e.g., CEO digital twin) or COO (e.g., COO digital twin); Alternatively, it may interact with and share such cost or budget data and reporting with other enterprise twins as described herein, including other suitable enterprise digital twins. In embodiments, CFO digital twin 8304 may utilize data analytics, machine learning, and A.I., as described herein, to provide financial reporting, projections, simulations, budgets, and related summaries. One or more intelligence services of EMP 8000 may be utilized based at least in part on the process. In some of these embodiments, CFO digital twin 8304 may be used to identify key departments, personnel, third parties or other persons that are listed in or subordinate to budget line items and therefore may have an interest in such materials, for example. Intelligent services can be used for this purpose. Budget material for a given party can be abstracted and summarized for presentation independently from the budget as a whole, formatted and presented automatically to the party that is the source of the cost and/or subject matter of the budget item, or at the direction of the CFO or other user. It can be.

In some embodiments, CFO digital twin 8304 may be configured to track and report inbound and outbound claims (i.e., accounts receivable and payables) associated with the finance department and/or organization. In an embodiment, CFO digital twin 8304 may include a billing digital twin that identifies billing departments, people, processes, and systems associated with an enterprise's billing workflow. In such embodiments, the billing digital twin may interact with the finance department to present, store, analyze, reconcile and/or report on billing activity related to the parties with which it is interacting. In some embodiments, a user of CFO digital twin 8304 may, if CFO digital twin 8304, approve an invoice via a GUI, issue an invoice, drill down into a set of invoices, initiate an investigation of invoices, etc. can be performed.

In embodiments, CFO digital twin 8304 may be configured to provide information unique to CFO digital twin 8304 to a user (e.g., a CFO or other finance department executive), and thus CFO digital twin 8304 can provide unique insights and perspectives on financial performance. For example, in supply chain planning, demand forecasting, operational planning, and other CFO activities, traditional data sources, models, and projections may be quantitatively robust within specific domains. Although the domain may be "siloed" in a way that means that the data can be However, it may be limited by factors other than these. In embodiments, EMP 8000 associated with CFO digital twin 8304 includes, but is not limited to, functions such as native data and model generation, and data and model combination and aggregation based at least in part on the real-time operations of the organization. Create and derive new financial metrics and analyses. Native data and model creation, such as specifying the data to be collected, the format for collecting and storing the data, the transformation of the data into a model, etc., rather than relying on data and/or model presets, can be implemented using EMP (8000) and CFO digital twins ( Create, combine, aggregate, modify native data (including combinations with other third-party data) in a manner that is mathematically tuned as appropriate for modeling, analysis, machine learning, and/or AI techniques performed by 8304); Provides transformation and/or weighting capabilities. Similarly, in the analytical context of CFO's behavior and functionality of EMP and CFO digital twins (8304), native data and model creation and structuring by EMP and CFO digital twins (8304) enable analytics, machine learning, AI behavior, etc. generates new analytical results and insights, based at least in part on the real-time behavior of the organization, because EMP and CFO digital twin 8304 enables CFOs to move further upstream in their financial data generation and modeling operations. Develop analytical insights that can be generated and reported for the purpose of improving performance, including but not limited to product margins (e.g., gross, contribution, net, etc.), product features, upsell opportunities, or some other performance metric. This is because it allows you to assert greater creative control over the types of data and other input materials that will be used for processing.

In embodiments, CFO digital twin 8304 may be configured to simulate finance-related activities on behalf of a user. In these embodiments, users may set financial and/or budget parameters, set price and sales goals, design processes, and maintenance/infrastructure upgrades, design internal controls, frequency/type of product testing, manufacturing downtime, flexible workforce planning, etc. One or more parameters may be identified that may vary during simulation, including but not limited to: In this embodiment, digital twin simulation system 8116 may receive a request to perform a simulation requested by CFO digital twin 8304, where the request indicates characteristics and parameters, including financial parameters, to be changed. In response, digital twin simulation system 8116 may return simulation results to CFO digital twin 8304, which in turn outputs the results to the user via a client device display. In this way, the user is provided with a variety of results corresponding to different parameter configurations. In some embodiments, a user can select a set of parameters based on various results. In some embodiments, an executive agent trained by the user may select a set of parameters based on various results. Simulations, analyzes and/or modeling performed by CFO Digital Twin 8304 can be used to mitigate risk for IPOs, M&A, capital raising and bond issuance, or some other type of transaction. Simulations, analysis and/or modeling performed by CFO digital twin 8304 may be used to create and structure sales incentives, including commissions and other performance-based rewards. Simulations, analyzes and/or modeling performed by CFO digital twin 8304 may be used to evaluate insurance offerings and other information related to business interruption preparedness. Simulations, analysis and/or modeling performed by CFO digital twin 8304 may be used to monitor loan commitments and analyze projections. Equipped with a digital twin (8304), CFOs will be better able to adapt quickly to change by anticipating headwinds, anticipating operational performance, and making informed decisions across departments while mitigating risk.

In embodiments, CFO digital twin 8304 may be configured to manage operating plans based at least in part by utilizing predictive analytics for sales planning and supply chain management to increase company efficiency while optimizing operating costs.

In an embodiment, CFO digital twin 8304 drives enumerated actions to a repository while providing insight across environmental resource management (ERM) solutions for risk oversight, including but not limited to internal control design, testing, certification, and reporting. Can be configured to access. In embodiments, CFO digital twin 8304 may be configured to streamline governance, risk management, and compliance processes to link risk and compliance across an organization and manage complex audit fieldwork and work documents.

In embodiments, CFO digital twin 8304 may be configured to store, aggregate, merge, analyze, prepare, report, and distribute data related to financial strategies, plans, activities, or initiatives. For example, CFO digital twin 8304 may be a plurality of financial data, summaries and reports and analyzes including financial data, summaries and reports and analyzes related to prior financial activities (e.g., prior quarterly financial performance). may be associated with a database or other repository, each of which may further be associated with third party financial or economic data.

In embodiments, CFO digital twin 8304 is configured to store, aggregate, merge, analyze, prepare, report, and distribute material related to financial reporting, valuations, rankings, financial trend data, earnings data, or other finance department related data. It can be. CFO Digital Twin 8304 is capable of linking to, interacting with, and associating with external data sources, uploading, downloading, and aggregating external data sources, including internal data of the EMP, and analyzing such data. You can. Data analytics, machine learning, AI processing, and other data-driven processes may be coordinated between CFO digital twin 8304 and the analytics team based at least in part on insights derived by artificial intelligence service system 8010. These collaborations and interactions seed finance-related data elements and domains into the enterprise data repository 8012 for use in modeling, machine learning, and AI processing to identify optimal financial strategies, or some other finance-related metric or aspect. In addition, it may also include assisting in the identification of optimal data measurement parameters upon which to base judgments on the success of financial efforts. Examples of data sources 8020 that may be connected to, associated with, and/or accessible from CFO digital twin 8304 include sensor system 8022, sales database 8024 that is updated with sales figures in real time; A CRM system (8026), a news website (8048), a financial database (8030) that tracks the costs of the business, an organizational chart (8034), a workflow management system (8036), a customer database (1S40) that stores customer data, and /or may include, but is not limited to, a third party data source 8038 that stores third party data.

In embodiments, CFO digital twin 8304 may aggregate data sources and types to create new data types, summaries, and reports that are not available elsewhere. This can reduce dependency on multiple third-party providers and the need for current solutions. This can reduce the costs associated with obtaining the data needed for sound financial decisions, among other benefits and improvements.

In embodiments, CFO digital twin 8304 may be configured to monitor a user's performance of finance-related tasks through the monitoring functionality of an agent of a client application 8052 executing CFO digital twin 8304. In embodiments, the executive agent's monitoring function may report to EMP 8000 about certain activities performed by users when interfacing with CFO digital twin 8304. In response, EMP 8000 may train an execution agent (which may include one or more machine learning models) to handle and process such finance-related tasks when they next arrive. For example, the monitoring function may be performed when a user (e.g., CFO) escalates the status of CFO digital twin 8304 to the CEO and/or when a user delegates a task to a subordinate via CFO digital twin 8304. can be monitored. Actions taken by you in response to each of your respective accounts whenever such evacuation and/or delegation events occur and/or when you (e.g., CFO or other financial executive) responds to an alert or other notification of an urgent nature; can be reported and can be reported to EMP (8000). In response, expert agent system 8008 may train executive agent 8364 based on the reported actions, which in turn may train executive agent 8364 to determine the specific later occurring event for which it was trained (e.g., It can be leveraged by the CFO digital twin to respond to financial performance or analytics showing financial activity (e.g. new investments). For example, a trained executive agent 8364 in conjunction with CFO digital twin 8304 may automatically issue financial performance alerts to specific employees based on performance trends of one or more business units. In another example, execution agent 8304 may automatically escalate a notification to the CEO (which may be displayed in CEO digital twin 8302) when certain metrics indicate poor financial projections. In embodiments, the executive agent 8364 associated with the CFO digital twin 8304 may allow the CFO to optimize initiatives in real time without the need to manually request such real-time financial performance data. In some embodiments, CFO digital twin 8304 may automatically present such information and relevant/necessary alerts as configured by the configuring user, CFO, or some other user with such permission.

In embodiments, trained executive agents 8364 in connection with CFO digital twin 8304 may, within the CFO's responsibilities, monitor the finance department's performance, finance department's personnel, finance activities, finance content, finance platform, finance partners, or management. may be configured to report on some other aspect of . Reports may be to the CEO, board of directors, other executives of the organization (e.g., COO), or external third parties (e.g., partners, press releases, etc.). The reporting capabilities of the CFO Digital Twin (8304) can also be used to populate the data needed for formal reporting requirements such as shareholder statements, annual reports, SEC filings, etc. Templates of common reporting formats can be stored and associated with the CFO digital twin (8304) to automate the presentation of data and analysis according to predefined formats, styles, and system requirements.

In embodiments, CFO digital twin 8304 in combination with EMP 8000 is configured to monitor, store, aggregate, merge, analyze, prepare, report, and distribute material related to competitors of the CFO's organization, or named entities of interest. It can be configured. In embodiments, such data may be aggregated, spidered, or used to retrieve and gather competitor information from sources including, but not limited to, press releases, SEC or other financial reporting, merger and acquisition activity, or some other publicly available data. It may be collected by EMP 8000 through ringing, web-scraping, or other techniques.

In embodiments, CFO digital twin 8304, in combination with EMP 8000, may monitor, store, aggregate, merge, analyze, prepare, and/or store material related to regulatory activities, such as government regulations, industry best practices, or some other requirements or standards. Can be configured to report and distribute. For example, the CFO digital twin (8304) can communicate with another enterprise digital twin, such as the general counsel digital twin (8314), which allows the legal team to keep the CFO informed of new regulations or regulatory changes as they occur.

In an embodiment, a client application 8052 running a CFO digital twin 8304 may relay the CFO's behaviors and preferences (or the behaviors and preferences of other financial personnel) to the expert agent system 8008, as described herein. The expert agent system 8008 may be configured to determine how a CFO or other financial personnel would respond to a particular situation and conduct data collection, analysis, machine learning, and A.I., as described herein. An executive agent can be trained to coordinate its actions based at least in part on technology. The foregoing examples are optional examples and are not intended to limit the scope of the present disclosure.

In this example of the Finance Department and CFO Digital Twin 8304, references to the features and functionality of the EMP and Digital Twin refer to other departments and Digital Twins, and their respective projects and It should be understood as applying to the workflow.

In embodiments, Chief Operating Officer (COO) digital twin 8306 may be a digital twin configured for an enterprise's COO or similar executive responsible for overseeing the enterprise's operational tasks. COO Digital Twin (8306) is responsible for the management of personnel and partners, supervision of various departments (e.g., supervision of marketing department, HR department, sales department, etc.), project management, implementation and/or rollout of business processes and workflows, May provide functions including, but not limited to, budgeting, reporting, and many other operational-related tasks.

In embodiments, COO digital twin 8306 may provide real-time, historical, aggregated, comparative, and/or projected financial information (e.g., sales, expenses, revenue, debt, profitability, cash flow, etc.), merger and acquisition information, , may provide data, analysis, summaries, and/or reports, including but not limited to system data, reporting and control data, or some other operation-related information. In embodiments, COO digital twin 8306 operates in conjunction with EMP 8000 to perform simulations, predictions, statistical summaries, analytics, machine learning, and/or other AI-based decision support and input (e.g., equipment data , sensor data, etc.), for example, those related to the development, communication and implementation of effective growth strategies and processes for the organization.

In embodiments, the types of data that may populate a COO digital twin may include, but are not limited to: operational data, key performance indicators (KPIs) for plants/plants, business units, and assets/equipment. ; Uptime/downtime, safety data, risk management data, supply chain/component availability data, demand planning data, logistics data, workflow data, by business unit, by product, by geography, by factory, by store location(s), Financial performance metrics by asset class, revenue, and resource utilization; Audit data, asset performance data, asset aging and depreciation data, asset allocation data, or some other type of operational-related data or information.

In embodiments, COO digital twin 8306 identifies, assigns, directs, supervises, and reviews operational department personnel and third-party personnel associated with the design, implementation, and evaluation of operational processes, internal infrastructure, reporting systems, company policies, etc. To do this, you can display twins of operating departments that the user can use.

In an embodiment, COO digital twin 8306 may include definitions of the various roles/employees working under the COO, reporting structures, and associated permissions for each individual within the business unit, and of the individuals filling each role. It can be filled with various names and/or other identifiers.

In embodiments, COO digital twin 8306 may be configured to interface with collaboration suite 8006 to specify and provide a set of collaboration tools that operational departments and associated parties can utilize. Collaboration tools may include video conferencing tools, “in-twin” collaboration tools, whiteboard tools, presentation tools, word processing tools, spreadsheet tools, etc., as described herein. Collaboration and communication rules may be constructed based at least in part on using an AI reporting tool, as described herein.

In some of these embodiments, COO digital twin 8306 may be configured to simulate operational activities, such as proposed new operational plans, processes, or programs. In such embodiments, the digital twin simulation system 8116 may receive a request to perform a simulation requested by the COO digital twin 8306, where the request includes features of the action plan or other activity proposed for implementation, and Representing parameters, and their associated variables, can be altered or changed to create different simulation environments. In response, the digital twin simulation system 8116 may return simulation results to the COO digital twin 8306, which in turn outputs the results to the user through the client device display. In this way, the user is provided with a variety of results corresponding to different operating parameter configurations. In embodiments, an executive agent trained by a user may select a set of parameters based on various results.

In embodiments, COO digital twin 8306 may be configured to store, aggregate, merge, analyze, prepare, report, and distribute material related to operational strategies, plans, activities, or initiatives. For example, COO digital twin 8306 may be associated with a plurality of databases or other repositories of operational data, summaries and reports and analyzes, each of which contains such materials, summaries and reports and analyzes related to previous operational activities. may be further associated with financial and performance metrics related to the activity and is also accessible to the COO digital twin 8306.

In an embodiment, the COO digital twin 8306 operates, including in real time, based at least in part on the use of a monitoring agent in a client application 8052 associated with the COO digital twin 8306, as described herein. Can be configured to monitor performance. The monitoring agent may report these activities to EMP 8000 for presentation in a user interface associated with COO digital twin 8306. In response, EMP 8000 (which may include one or more machine learned models) handles and processes these notifications the next time they arrive and escalates and/or alerts the COO when these notifications are of an urgent nature. ) can train enforcement agents.

In embodiments, COO digital twin 8306 may be configured to report on operational unit performance, operational unit personnel, operational activities, operational content, operational platforms, operational partners, or some other aspect of management within the COO's responsibility. there is.

In embodiments, EMP 100 trains and deploys enforcement agents on behalf of enterprise users. In embodiments, an executive agent is an AI-based software system that performs tasks on behalf of each executive user and/or suggests actions to each executive user. In embodiments, EMP 100 receives data from various data sources associated with a particular entity or workflow and learns the workflow performed by a particular user based on the data and surrounding situation or context. For example, the user may be a COO for whom COO digital twin 8306 is presented. Among the COO's responsibilities may be scheduling maintenance and replacement of equipment in a manufacturing, warehouse, or other operating facility. The status displayed in COO digital twin 8306 may include an indication of the condition of different equipment within the operating facility. In this example, the COO may schedule maintenance via the digital twin when it is determined that a piece of equipment is in a first condition (e.g., a deteriorating condition), and the COO can schedule maintenance when the piece of equipment is determined to be in a first condition (e.g., a critical condition). ), a request to replace the equipment can be issued to the COO via the COO digital twin (8306). The executive agent may learn the COO's tendencies based on the COO's previous interactions with the COO digital twin 8306. Once trained, the execution agent can automatically request a replacement from the COO when a particular piece of equipment is determined to be in a second condition, and can automatically schedule maintenance if the piece of equipment is in a first condition.

In an embodiment, a client application 8052 running a COO digital twin 8306 may relay the COO's behaviors and preferences (or the behaviors and preferences of other operational personnel) to the executive agent system 8008, as described herein. Execution agent system 8008 may be configured to determine how the COO or other executive personnel respond to a particular situation and perform data collection, analysis, machine learning and A.I., as described herein. An executive agent can be trained to coordinate its actions based at least in part on technology. The foregoing examples are optional examples and are not intended to limit the scope of the present disclosure.

In this example of the Operations Department and COO Digital Twin (8306), references to the features and functionality of the EMP and Digital Twin are to other departments and Digital Twins, and their respective projects and tasks, unless the context indicates otherwise. It should be understood that it also applies to flow.

In embodiments, Chief Marketing Officer (CMO) digital twin 8308 may be a digital twin configured for an enterprise's CMO, or similar executive responsible for overseeing the enterprise's marketing tasks. CMO Digital Twin (8308) manages personnel and partners, develops and supervises marketing budgets and resources, manages marketing and advertising platforms, develops and manages marketing content, strategies and campaigns, reporting, competitor analysis, regulatory analysis, and data. May provide functionality including, but not limited to, management of privacy and security.

In embodiments, types of data that may be populated and/or utilized by CMO digital twin 8308 may include, but are not limited to: macroeconomic data; market price data; Competitive product and pricing data; microeconomic analysis data; expected data; demand planning data; competition matrix data; product roadmap; product capability data; consumer behavior data; consumer profile data; collaborative filtering data; Analytical results from AI and/or machine learning modeling; channel data; demographic data; geographic data; predictive data; Referral data, or some other type of data related to the operations of the CMO and/or marketing department.

In embodiments, an executive digital twin, such as CMO digital twin 8308 or another executive digital twin, may represent the twin of a department, such as a marketing department or another department, and the user may identify third-party partners involved in the organization's related efforts. It may be used to identify, assign, direct, supervise and review departmental personnel and third-party personnel associated with the activities of specific departments of the organization, including other external colleagues. Examples of these organizational personnel include the organization's marketing staff, sales staff, finance staff, product design staff, engineers, analysts, statisticians, data scientists, advertising staff, executive staff, human resources staff, board members, advisors, or some other type of staff. Including, but not limited to, organizational personnel. Examples of your organization's third-party personnel include advertising agency employees, ad exchange employees, external creative or content developers, social media platform personnel, joint marketing partners, consultants, contractors, finance firm personnel, auditors, or some other type of third-party personnel. Including, but not limited to. In an embodiment, a department twin (in this example, a marketing department twin) includes definitions of the various roles/employees working under an executive (e.g., CMO), reporting structures, and associated permissions for each individual within the business unit. may be populated with various names and/or other identifiers of the individuals filling each role. In embodiments, a department twin (eg, a marketing department twin) may include subsections specific to activities or initiatives, such as marketing or advertising campaigns. In this way, executives (e.g., CMOs) can easily identify personnel and third-party providers involved in an initiative and/or assign individuals and/or third parties to the initiative. Users may access business information (e.g., using enterprise configuration system 8002), as described herein, such that restrictions, permissions, and/or access rights may be controlled by the CMO (or similar user). You may define one or more restrictions, permissions, and/or access rights for individuals represented in the unit. In embodiments, permissions to define such restrictions and/or permissions may be provided, for example, by an organization listing users as having roles that allow it to implement permissions, restrictions, and/or access permissions for roles/individuals. Can be defined in digital twin. In embodiments, personnel restrictions or rights associated with a role/individual may be specific to a project, such as a marketing or advertising campaign, and may or may not be permitted to access (either directly or in a digital twin) certain users or groups of users. More types of data can be defined. For example, a primary marketing campaign twin could allow marketing department staff to review the primary marketing budget for a primary marketing campaign and approve marketing spending for the primary marketing campaign up to $10,000, but not for the secondary marketing campaign. Twins can prevent the same staff from doing any budget review or spending. Similar approaches can be used by various types of projects across organizations and departments, such as product development projects, logistics projects, company development projects, service projects, etc. In embodiments, a violation or attempted violation of a restriction, permit, or access right may invoke a notice, alert, warning, or some other action that notifies an individual of the violation or attempted violation. In examples, such notifications, alerts, or warnings may be sent to an identified individual based at least in part on the individual's position in the organizational chart regarding the person violating or attempting to violate a restriction, permit, or access right. In other examples, such notifications, alerts, or alerts may be sent to individuals who are not identified in the departmental org chart and/or in a specific project or campaign, but rather are identified based at least in part on rules defined in the overall enterprise's organizational twin. It can be sent to individuals. For example, rules stored within an entity's organizational digital twin may specify that an alert should be sent to an information security department staff member or some other staff member when an attempt is made to log into a prohibited file or other system. Other rules may relate to geographic, temporal, or other types of restrictions, as described herein. In embodiments, the alert may be an email, phone call, text, or some other type of communication.

In embodiments, CMO digital twin 8308 may be configured to oversee and manage staffing and personnel issues and activities related to the marketing department. For example, a marketing department twin could map each individual within the marketing department to their respective marketing department. Using the CMO Digital Twin (8308), users can select a department to see more details about the department's functions. Alternatively, this step may be performed automatically by the CMO digital twin 8308, requiring no action from the user (e.g., CMO) (e.g., via an enforcement agent trained by the user). I never do that. For example, further details may include the number of vacancies currently associated with the department and how long each open position will remain unfilled, estimated salary data associated with the open positions, etc. The user may also choose to view more information about the budget associated with a given department, such as which department has staff vacancies, to see if there is currently available budget to cover new hires for the department. Alternatively, this step can be performed automatically by the CMO digital twin 8308, requiring no action from the user. Continuing the example, if there is a budget to cover new hires, the CMO digital twin 8308 provides a link for the user to initiate communication with Human Resources or some other department personnel to begin the process of posting a job listing. Or it may provide another opportunity. Alternatively, this step may be performed automatically by CMO digital twin 8308 (e.g., through an enforcement agent executing on behalf of the user), requiring no action from the user. This communication may be retrieved from a repository of emails, letters, or other forms of communication such that the user does not need to compose the communication, but instead simply signals within the CMO digital twin 8308 that such communication should be sent. Similarly, based on the communication type (e.g., “initiating a new marketing task posting”), the user needs to select a receiving party that can be stored in the EMP as an appropriate recipient based at least in part on rules associated with the communication type. There may be no. Continuing with additional examples, alternatively, if there is no budget available to cover new hires, a second type of communication may be invoked by CMO digital twin 8308, e.g., email, meetings, etc. A calendar invitation to schedule a meeting, or some other type of communication, may be selected to be sent to the CFO or other financial personnel, requesting a meeting to discuss the marketing department's budget or initiate some other activity. According to this example, if and when a new hire is approved, the CMO digital twin may allow the user to delegate hiring tasks to a subordinate. If a user is assigned to hire a new employee, CMO digital twin 8308 can provide data about the candidate (e.g., resumes, recommendations, interview notes from interviewer, etc.), and the user can consider or , you can select one or more candidates to interview or hire.

In one example, a user can select a subdepartment within the marketing department to view the subdepartment's performance in greater detail. For example, further details could include the number of types of training sessions selected marketing department employees receive, tutorials, events, conferences, etc. Users can compare these training and event attendance levels to specified target criteria stored in or associated with the EMP. This may cause CMO digital twin 8308 to report to the CMO a list of employees within her department whose training and/or event attendance does not meet target criteria. This list can be prioritized by the CMO digital twin (8308) to highlight staff members most in need of additional training. Users can also obtain more information about the budget associated with a given department, such as which department has staff that do not have adequate training according to target criteria, to see if there is currently available budget to cover additional training for the department. You can choose to view it. If there is a budget to cover additional training, CMO digital twin 8308 communicates to staff members who need training, for example, to alert them that they should schedule training and/or event attendance within a time frame. You may provide the user with a link or other opportunity to initiate. This communication may be retrieved from a repository of emails, letters, or other forms of communication such that the user does not need to compose the communication, but instead simply signals within the CMO digital twin 8308 that such communication should be sent. Continuing the example further, a second type of communication may be invoked by the CMO digital twin 8308, such as a request for information, training registration, or some other type of communication scheduling training and/or events. It may be chosen to be sent to a third party training vendor used by a marketing department, conference event registration, or other training or event entity to request registration, or some other activity. Alternatively, the steps discussed above for tracking and reporting on marketing staff training and attendance can be performed automatically by CMO digital twin 8308, requiring no action from the user. In this example of the Marketing Department and CMO Digital Twin (8308), references to the features and functionality of the EMP and Digital Twin refer to other departments and Digital Twins, and their respective projects and It should be understood as applying to the workflow.

In embodiments, CMO digital twin 8308 may be configured to interface with collaboration suite 8006 to specify and provide a set of collaboration tools that the marketing department and associated parties can utilize. Collaboration tools may include video conferencing tools, “in-twin” collaboration tools, whiteboard tools, presentation tools, word processing tools, spreadsheet tools, etc., as described herein. Collaboration and communication rules may be constructed based at least in part on using an AI reporting tool, as described herein.

In embodiments, CMO digital twin 8308 may support a marketing department, including, but not limited to, an entire department budget, a budget for a single or group of marketing or advertising campaigns, a budget for a third-party vendor, or some other type of budget. It can be configured to research, create, track and report on budgets. CMO digital twin 8308 interacts with other executive twins as described in this application, including, but not limited to, digital twins associated with the finance department, accounts payable, executive staff such as the CEO and CFO, etc., and compiles such budget data and reporting. You can share it. CMO Digital Twin 8308, as described herein, provides data analytics, machine learning, and A.I. Marketing budgets and related summaries, based at least in part on the process, for example, to identify key departments, employees, third parties or other persons listed in or subordinate to budget line items and who may therefore have an interest in such materials; and intelligence that reads data. Budget data for a given party can be abstracted and summarized for presentation independently from the budget as a whole, and formatted and presented automatically or at the user's direction to the party that is the subject of the budget item. In a simplified example, the CMO might create a new marketing campaign, “Airlines - Airfare Coupon Text Campaign - January,” that includes the following line items: $15,000 for third-party advertising agency content creation; Social Media Platform Placement $50,000; Analytics Department $25,000, etc. The overall budget can be shared (at the user's option or automatically) with parties who need to approve the entire budget, such as the CFO. As described in this application, this sharing can be accomplished by the CMO digital twin 8308 communicating directly with the CFO digital twin, such that the information is shared with the CFO without requiring the CFO to have knowledge of the budget or requiring the CFO to know the budget. is presented in Subparts of the budget, for example, analytics department line items, are automatically sent by the CMO digital twin (8308) to the head of the analytics department to report the total amount of authorized spend approved for that department for a specific marketing campaign. You can notify.

In embodiments, CMO digital twin 8308 may be configured to track and report inbound and outbound claims (i.e., accounts receivable and accounts payable) associated with the marketing department. Billing departments, employees, processes and systems, including the billing digital twin, interact with the CMO digital twin (8308) to communicate with advertising agencies, advertising networks, ad exchanges, content creators, advertisers, social media platforms, television, radio, online entities, etc. Likewise, the marketing department may present, store, analyze, reconcile and/or report billing activity related to the parties with which it is contracting.

In an embodiment, CMO digital twin 8308 may be configured to display a marketing campaign twin. In this embodiment, CMO digital twin 8308 may store marketing content associated with a marketing campaign, market research conducted in connection with the marketing campaign, tracking data of the marketing content associated with the marketing campaign (e.g., geographic reach of the marketing campaign, may display various statuses and/or items associated with a marking campaign, such as demographic data associated with the campaign, analytics of the marketing campaign (e.g., results associated with the marketing campaign on various platforms), etc. In some embodiments, the CMO digital twin may be configured to automatically report on marketing campaign-related activities through a user interface associated with CMO digital twin 8308. These activities can be determined using marketing department metadata that indicates status changes such as changes to website content, changes to product photos in advertisements, changes to wording in mailings, etc. CMO Digital Twin 8308 may also display activity among classes of entities that are monitored or designated for monitoring in CMO Digital Twin 8308, such as a new press release regarding a discounted advertising opportunity available from an advertising exchange. In embodiments, CMO digital twin 8308 provides research, tracking, monitoring, and analysis of media content performance across a variety of marketing-related platforms, and automatically reports such activities to a user interface associated with CMO digital twin 8308. Can be configured to report. These platforms may include, but are not limited to, customer relationship platforms (CRMs), organizational website(s), social media, blogs, press releases, mailings, in-store or other promotions, or some other type of marketing platform-related materials or activities. , but is not limited to these.

In some of these embodiments, CMO digital twin 8308 may be configured to simulate a marketing campaign, such that the simulation of a marketing campaign can include vehicle (e.g., social media, television, billboard, print, etc.), budget, targeting parameters, etc. (e.g., geographic, demographic, etc.), and/or other appropriate marketing campaign parameters. In this embodiment, digital twin simulation system 8116 may receive a request to perform a simulated CMO digital twin, where the request indicates campaign characteristics and parameters to be changed. In response, digital twin simulation 8116 may return simulation results to CMO digital twin 8308, which in turn outputs the results to the user through the client device display. In this way, the user is provided with a variety of results corresponding to different parameter configurations. In some embodiments, a user can select a set of parameters based on various results. In some embodiments, an executive agent trained by the user may select a set of parameters based on various results.

In embodiments, CMO digital twin 8308 may be configured to store, aggregate, merge, analyze, prepare, report, and distribute material related to marketing strategies, plans, campaigns, or initiatives. For example, CMO digital twin 8308 may be associated with a plurality of databases or other repositories of marketing presentation materials, summaries and reports and analyzes including such presentation materials, summaries and reports and analyzes related to previous marketing campaigns; Each of these may be further associated with financial and performance metrics related to the campaign and are also accessible to CMO digital twin 8308. Such historical marketing campaign material may consist of advertising, marketing or other content that may be categorized based in part on the financial and performance metrics to which it is associated. For example, there may be a first category referred to as “Market Tested Content,” which consists of content that is fielded in marketing campaigns within customer segments, so that its actual performance is based on actual market testing. It is completely known. Because marketing content from this category has been field tested, the content may be scored based at least in part on the financial, performance, or other data to which it is associated. A second category could be “New Content - Simulation Tested”, which is not fielded, but includes simulated customer segmentation analysis, simulated A/B testing, simulated attribute modeling, simulation A.I., including but not limited to market mix modeling, machine learning, classification, probabilistic modeling, learning techniques, etc. It consists of content to which analytical tests such as technology have been applied. Because the marketing content from this category has been simulation tested, the content may be scored based at least in part on simulated performance data or other data to which it is associated. To continue with the example, the third category of content is "New Content - Panel Tested", which has not been fielded or simulated, but whose views, opinions and impressions are shared between human panels. It consists of content tested in . Because marketing content from this category has been human panel tested, the content may be scored based at least in part on performance data as reported by the human panel, or other data to which it is associated. A final fourth category of content may be “New - Untested,” which may be newly developed or other content that has not been tested in the field, in simulations, or by a human panel. CMO Digital Twin 8308, as described in this application, uses machine learning, A.I. and other analytics capabilities, to analyze the content of four categories of content and classify and score content characteristics that are probabilistically associated with improved financial or other performance for a stated type of marketing campaign or marketing topic. there is. Statistical weights may be applied to these characteristics, where the weights represent a greater degree of financial or some performance metric of interest. Similarly, market characteristics may be analyzed for marketing content to determine consumer characteristics that are probabilistically associated with improved financial or other performance for a given marketing content. CMO digital twin 8308 can provide a user interface that enables access to this repository of stored data about content categories, consumers, and performance. When planning a marketing campaign, a CMO or other marketing personnel can use the CMO digital twin 8308 to select from this repository of content content that will probabilistically better serve as the intended consumer target for the new campaign. For example, from past marketing field tests from actual previous marketing campaigns, data showed that marketing content with images of large dogs outperformed content with pictures of small dogs (e.g., based on ad conversion rates), and that this effect was influenced by age and amount. (i.e., older adults have a significantly greater preference for larger dogs). Performance data from simulation-tested content may show a similar but smaller effect based on the size of the dog image in the content, while panel-tested data may show a similar effect for larger dog images in the content. , one may also have performance data showing, based on panel data, that the effect appears to diminish for people under 15 years of age (i.e., younger people are more attracted to smaller dog breeds than older people). For CMOs using the CMO Digital Twin (8308), this data, and the more successful characteristics of the content, will be most effective in new marketing campaigns intended to sell soft drinks from the fourth category of content (“New - Untested”). It can be used to select appropriate content. In embodiments, the artificial intelligence service system 8010 of the EMP 8000 may select content and segment its presentation based at least in part on previous performance data, such that it can be used on platforms that tend to have people over the age of 15. The advertising presented will use content predominant with large breed dogs, while platforms with younger audiences will offer a preference for a larger mix of dog breeds and possibly small breed dogs in marketing images. When a marketing campaign is deployed in the field, the CMO digital twin 8308 can monitor, track, and report on the performance of the marketing campaign so that the CMO can review and intervene as needed. Once new content has been field tested, it can be stored and categorized in the first category of content, “market tested content,” along with associated financial and performance metrics. In other examples, similar stored content, content categories, characteristics, and financial and performance metrics may be used by CMO digital twin 8308 to recommend, for example, search engine optimization (SEO), or other marketing strategies and techniques. You can.

In embodiments, CMO digital twin 8308 stores, aggregates, merges, analyzes, prepares, and reports material related to market surveys, online surveys, customer panels, ratings, rankings, marketing trend data, or other data related to marketing. and may be configured to distribute. CMO digital twin 8308 is capable of linking to, interacting with, and associating with external data sources and uploading, downloading, and executing external data sources, including internal data of the EMP, as described herein. You can aggregate and analyze this data. Data analytics, machine learning, AI processing, and other analytics may be coordinated between CMO digital twin 8308 and the analytics team based at least in part on using artificial intelligence service system 8010. These collaborations and interactions can be used in enterprise data repositories (such as corporate data repositories) for use in modeling, machine learning, and AI processing to identify optimal marketing content, sales channels, target consumers, price points, timing, or some other marketing-related metric or aspect. 8012), but may also include assisting in the identification of optimal data measurement parameters upon which to base judgments on the success of marketing efforts. Examples of data sources 8020 that can be connected to, associated with, and/or accessed from the CMO digital twin 8308 include a sensor system 8022, a sales database 8024 that is updated with sales figures in real time; CRM systems (8026), content marketing platforms (8028), news websites, financial databases that track the costs of a business (8030), surveys (8032) (e.g., customer satisfaction surveys), organizational charts (8034), It may include, but is not limited to, a workflow management system 8036, a customer database 8040 to store customer data, and/or a third-party data source 8038 to store third-party data.

In embodiments, CMO digital twin 8308 may be configured to assist in the development of new marketing campaigns. For example, CMO digital twin 8308 can identify internal and external partner teams for a marketing campaign. For example, individuals who are ideal candidates to assist with a marketing campaign may be identified based at least in part on experience and expertise data stored within or in association with CMO digital twin 8308. In another example, CMO digital twin 8308 can identify marketing campaign objectives, record, monitor, and track the performance of the campaigns against those objectives, and track the campaigns within a user interface associated with CMO digital twin 8308. can be presented in real time. Examples of marketing targets include unit distribution, customer acquisition, customer retention, customer chum, customer loyalty (e.g., repeat purchases), customer acquisition cost, duration of average sales cycle, advertising conversion rate, sales growth, and sales volume. Includes, but is not limited to, geographic expansion, demographic expansion of sales, market penetration, percentage of market control, marketing campaign ROI, regional comparison of performance, channel analysis, sales partner analysis, marketing partner analysis, or some other marketing target. No.

In an embodiment, CMO digital twin 8308 may provide a customer feedback loop, customer feedback, and feedback based at least in part on the use of a monitoring agent in client application 8052, as described herein, associated with CMO digital twin 8308. , can be configured to monitor customer satisfaction, complaints, product returns, etc. Such feedback data may include data derived from call center activity, chatbot activity, emails (e.g., complaints), product returns, Better Business Bureau submissions, or some other type of customer feedback or expression of customer opinion; It is not limited to this. Client application 8052 may include a monitoring agent that monitors how a customer or other person responds to a marketing campaign. The monitoring agent may report customer responses to these campaigns to EMP 8000 for presentation in a user interface associated with CMO digital twin 8308. In response, EMP 8000 trains an enforcement agent (which may include one or more machine learned models) to handle and process these notifications the next time they arrive, and determines whether these notifications are of an urgent nature, e.g. For example, the CMO may be notified and/or alerted upon the announcement of a class action lawsuit related to a product that is the subject of a marketing campaign. In embodiments, CMO digital twin 8308 may generate performance alerts based on performance trends. This can allow CMOs to optimize marketing campaigns in real time without having to manually request such real-time performance data; CMO digital twin 8308 can automatically present such information and relevant/necessary alerts as configured by the organization, CMO, or some other interested party.

In embodiments, CMO digital twin 8308 may be configured to report on marketing department performance, marketing department personnel, marketing campaigns, marketing content, marketing platforms, marketing partners, or some other aspect of management within the scope of the CMO. there is. Reports may be to the CMO, marketing department, other executives in the organization (e.g., CEO), or external third parties (e.g., marketing partners, press releases, etc.). As described herein, reporting may include sales summaries, customer data, marketing campaign performance metrics, cost per sale data, cost per conversion data, customer analytics such as predicted customer lifetime value for newly acquired customers, or some other May contain tangible reporting data. Reports and their contents may be shared by CMO digital twin 8308 with other executive digital twins, for example, data related to new customers, particularly those with high predicted customer lifetime value, for the purpose of exploring cross-selling opportunities. can be shared with sales staff. The reporting capabilities of CMO Digital Twin (8308) can also be used to populate the necessary data for formal reporting requirements such as shareholder statements, annual reports, SEC filings, etc. Templates of common reporting formats can be stored and associated with the CMO digital twin (8308) to automate the presentation of data and analysis according to predefined formats, styles and system requirements.

In embodiments, CMO digital twin 8308 may be configured to monitor, store, aggregate, merge, analyze, prepare, report, and distribute material related to the CMO's organization or competitors of the named entity of interest. In embodiments, such data may be aggregated, spidered, or used to retrieve and gather competitor information from sources including, but not limited to, press releases, SEC or other financial reporting, merger and acquisition activity, or some other publicly available data. It may be collected by EMP 8000 through ringing, web-scraping, or other techniques.

In embodiments, CMO digital twin 8308 may be configured to monitor, store, aggregate, merge, analyze, prepare, report, and distribute material related to regulatory activities, such as government regulations, industry best practices, or some other requirements or standards. there is. For example, the marketing industry is subject to data privacy and security laws in many jurisdictions, and this is an area of law and regulation that is experiencing rapid change. In embodiments, CMO digital twin 8308 may communicate with another enterprise digital twin, such as general counsel digital twin 8314, so that the legal department can keep the CMO informed when new regulations or regulatory changes occur. Similarly, as a CMO develops a new market campaign and selects the jurisdictions (e.g., United States vs. Europe) and populations (e.g., minors vs. adults) that will be part of the campaign, the CMO digital twin 8308 It may automatically transmit a synopsis of aspects of the campaign relevant to privacy law reviews so that they can be checked for legal and regulatory compliance prior to launch. In an example, such a marketing campaign synopsis may include the jurisdiction of the campaign, the intended audience, the means of obtaining consent, and the type of consent to be obtained (e.g., opt-in, opt-out, It may include a summary such as passive). Once approved and launched, as customer consent and other data privacy-related information is received by the organization, CMO digital twin 8308 allows the CMO to track metrics, such as future marketing materials (e.g., email requests). This may facilitate tracking the percentage of customers who choose to opt-in to receive. When the organization receives privacy-related material, it may, for example, receive a data subject request (DSR) from an EU citizen who has requested that their data be erased (i.e. who has exercised their “right to be forgotten”). ), we may store such information for future retrieval, summary, deletion or other actions. In embodiments, CMO digital twin 8308 monitors, stores, aggregates, merges, analyzes, prepares, reports, and reports information about what customer data is collected, who is responsible for collecting and storing it, the location and duration of storage, etc. It can be distributed. This data may be recalled by CMO digital twin 8308, for example, in the event of a data breach. The CMO digital twin 8308 can communicate with the CPO digital twin, including, for example, summarizing the list of people affected by the breach and the type of data breached and sharing this information with the Chief Privacy Officer (CPO). You can share it.

In an embodiment, a client application 8052 running a CMO digital twin 8308 may relay the CMO's behaviors and preferences (or the behaviors and preferences of other marketing personnel) to the expert agent system 8008, as described herein. The expert agent system 8008 may be configured to determine how a CMO or other marketing personnel would respond to a particular situation and conduct data collection, analysis, machine learning, and A.I., as described herein. An executive agent can be trained to coordinate its actions based at least in part on technology.

In embodiments, a Chief Technical Officer (CTO) digital twin 8310 is an enterprise's CTO or other technical executive responsible for overseeing and managing the enterprise's R&D, technology development, enterprise technology implementation, and/or engineering activities. It may be a digital twin constructed for this purpose. In an embodiment, CTO digital twin 8310 provides a real-time view of enterprise technology assets, including technology capabilities and versions. For example, in a manufacturing enterprise, CTO digital twin 8310 can indicate where environment-compatible updates, upgrades, or replacements may be available. CTO digital twin 8310 provides real-time, historical, aggregated, comparative, and/or projected technical information (e.g., real-time, historical, simulated, and/or projected technical performance data related to company products, benchmarking results, etc.) May provide data, analysis, summaries, and/or technical reporting, including but not limited to: CTOs using CTO Digital Twin (8310) can significantly better maintain technical development and software engineering impact by engaging in continuous virtualized learning using CTO Digital Twin (8310). In embodiments, the CTO digital twin 8310 may assist in virtual collaboration (CTO - an essential skill) because the CTO can use virtual tools to imagine and idealize to achieve something, often something that has not been done before. This is because the environment will require partnering with in-house engineers and external vendors. In embodiments, the CTO digital twin operates in conjunction with EMP 8000 to perform simulations, predictions, statistical summaries, analytics, machine learning, and/or other AI and inputs (e.g., technical performance data, sensor data, etc.). Can provide decision support based on learning-style processing.

In embodiments, CTO digital twin 8310 may be responsible for managing technical personnel, partners, and external consultants and contractors (e.g., developers, beta testers, etc.), overseeing budgets, procurement, spending, and policy compliance (e.g., code may provide features and functionality including, but not limited to, policies related to usage, storage, documentation, etc.), and other technology, development, and/or engineering-related resources, and/or reporting.

In embodiments, the types of data that can populate a CTO digital twin include technical performance and specification data, interoperability and compatibility data, cybersecurity data, competitor data, failure mode effects analysis (FMEA) data, technology/engineering roadmap data, and information. Technical system data (including relating to any of the hardware, software, networking, and other types mentioned or described in this application), operational technical and system data, uptime/downtime/operational performance data, asset aging/vintage. /Timing data, technical performance metrics by business unit, product, geography, factory, store location(s), resource utilization, competitive product and pricing data, forecasting data, demand planning data, analytics results from AI and/or Machine learning modeling (e.g., technical projections), forecast data, metrics regarding patent disclosures, patent applications, and/or patent grants, recommendation data, and/or the CTO and/or technology, development, and/or engineering departments. May include, but is not limited to, other types of data relevant to operations.

In embodiments, CTO digital twin 8310 may enable users to interact with the organization's Technology, Development, and/or Engineering Departments may be used to identify, assign, direct, supervise and review Technology, Development, and/or Engineering Department personnel and third party personnel associated with Technology, Development, and/or Engineering activities; A set of tweens can be displayed. Examples of such organizational personnel include, but are not limited to, technical, development, and/or engineering department employees, sales staff, and analysts, statisticians, data scientists, or some other type of organizational personnel associated with the functions of the technical, development, and/or engineering departments. Not limited. Examples of third party personnel in the technology, development, and/or engineering departments include management consultants, developers, software engineers, testers, and/or engineering partners, consultants, contractors, technology company employees, auditors, or some other type of third party. This includes, but is not limited to, personnel.

In an embodiment, CTO digital twin 8310 may include definitions of the various roles/employees working under the CTO, reporting structures, and associated permissions for each individual within the business unit, and of the individuals filling each role. It can be filled with various names and/or other identifiers.

In an embodiment, a client application 8052 running a CTO digital twin 8310 may use a collaboration suite 8006 to specify and provide a set of collaboration tools that technology, development, and/or engineering departments and associated parties can utilize. You can interface with Collaboration tools may include video conferencing tools, “in-twin” collaboration tools, whiteboard tools, presentation tools, word processing tools, spreadsheet tools, etc., as described herein. Collaboration and communication rules may be constructed based at least in part on using an AI reporting tool, as described herein. Collaboration and communication tools and associated rules can be configured to use company-specific, industry-specific, and domain-specific taxonomies and vocabularies when representing entities, states, and flows within the CTO digital twin 8310.

In embodiments, CTO digital twin 8310 may enable users to research, create, and track technology, development, and/or technology or engineering department initiatives, including but not limited to new product developments, updates, enhancements, replacements, upgrades, etc. and may be configured to enable reporting. In embodiments, CTO digital twin 8310 may associate and/or communicate with databases, including databases storing analytics and/or product data and product performance data, as described herein. Information can be presented on the interface associated with (8310). As product development progresses, real-time operations and other technical information may be used to continuously update product development summaries that can be used for review by the CTO or other technical personnel. CTO digital twin 8310 may also be associated with and/or communicate with databases, including databases storing analytical and/or competitive product data and product performance data, and may provide this information to the CTO, as described herein. It can be presented on the interface associated with the digital twin (8310). As the CTO's company's products change, and competitor products change, the current status and specifications can be presented by the CTO digital twin 8310 for the CTO or other technical personnel to review for direct product comparison. These comparisons may be used, in part, to generate analyzes, scores, reports, etc. that indicate the relative advantages and/or disadvantages that the company's product(s) have over competitor product(s). In an example, reports may be automatically provided to a marketing department to highlight the relative advantages (e.g., speed of processing) of a company's products over competitor products that should be used in new marketing campaigns. Sharing with the marketing department may be accomplished, in part, by CTO digital twin 8310 communicating with CMO digital twin 8308 to present reports or other information to the CMO or marketing staff.

In embodiments, CTO digital twin 8310 may be configured to present a simulation of technology development and/or engineering activities. For example, in some embodiments, digital twin system 8004 may simulate product use under multiple constraints that may affect product performance, such as operating environment, processing speed, storage, or other platform characteristics. In embodiments, real-time operational data, such as operational data available through EMP 100, may be included in the simulated data for the purpose of performing operational simulations. This allows the CTO to gain a deeper understanding of the behavior of the company's products in the real world and within modified, simulated real-world environments. It also enables simulated decision-making in a virtual environment and review of such digital twins in the context of the supplied solution and its relationship to the business, thereby aligning physical product production with business priorities to aid in the evaluation of vendor supplied solutions. It can allow for a product architecture based on interconnected operational digital twins to be built. In embodiments, simulations may also include technical and non-technical data in reporting, presentations, and dashboards for various technical and/or product specification parameters, product design and monitoring, internal control design, testing, certification, and technical decision making. It may include simulations related to the delivery of . In this embodiment, digital twin simulation system 8116 may receive a request to perform a simulation requested by CTO digital twin 8310, where the request indicates features and parameters, including technical parameters, to be changed. In response, the digital twin simulation system 81D16 may return simulation results to the CTO digital twin 8310, and the CTO digital twin in turn outputs the results to the user through the client device display. In this way, the user is provided with a variety of results corresponding to different technical and/or product parameter configurations. In some embodiments, a user can select a set of parameters based on various results. In some embodiments, an enforcement agent trained by the user may select a set of technical parameters based on various results. Simulation, analysis, and/or modeling performed by CTO digital twin 8310 may be used to reduce test time, design time, or some other type of technical cost. Simulation, analysis, and/or modeling performed by the CTO digital twin 8310 may be used to create and structure product development and test plans. Simulations, analysis, and/or modeling performed by CTO digital twin 8310 can be used to assess product-to-market timing and readiness. CTOs equipped with the CTO Digital Twin (8310) will be able to better adapt quickly to identify product and/or technical parameters requiring further development and predict the operational performance of the product. This can reduce errors, speed testing, reduce the need for patches, bug fixes, updates, etc. and smoothen agile process management.

In embodiments, CTO digital twin 8310 allows a user to include an entire department budget, a budget for a single or group of technology, development, and/or engineering initiatives, third-party vendor activities, or some other type of cost or budget. However, it may provide interfaces that enable research, creation, tracking, and reporting on, but not limited to, technical, development, and/or engineering department initiatives. CTO Digital Twin 8310 interacts with other executive twins, including, but not limited to, digital twins associated with accounts payable, executive staff such as the CEO, and/or others and shares such cost or budget data and reporting. can do.

In embodiments, CTO digital twin 8310 utilizes artificial intelligence service system 8010 (e.g., data analytics, machine learning, and A.I. processes) to read technical reports, projections, simulations, and related summaries and data. , for example, may identify key departments, employees, third parties or other persons listed in or subordinate to the technical item or detail provided.

In embodiments, CTO digital twin 8310 may be configured to provide information unique to CTO digital twin 8310 to the CTO, or other technical, development, and/or engineering department personnel, thereby enabling real-world and simulated activities. Both may provide insight and perspective on technical performance unique to the CTO Digital Twin 8310, based at least in part on the CTO Digital Twin 8310 utilizing real-time production, development and operational data.

In embodiments, CTO digital twin 8310 may be configured to manage operating plans based at least in part by utilizing predictive analytics for development planning and supply chain management to increase company efficiency while optimizing operating costs. In embodiments, CTO digital twin 8310 may be configured to obtain and display supervisory activities, including but not limited to internal control design, testing, and reporting, while directing the listed actions to appropriate personnel.

In embodiments, CTO digital twin 8310 may be configured to display, aggregate, merge, analyze, prepare, report, and distribute material related to technological strategies, plans, activities, or initiatives. For example, CTO digital twin 8310 may be a plurality of databases of knowledge bases, summaries and reports and analyzes including knowledge bases, summaries and reports and analyzes related to previous technical activities and results (e.g., bug testing) or It may be associated with other repositories, each of which may further be associated with third party technical or economic data, including competitor product data and/or technology benchmarks.

In embodiments, CTO digital twin 8310 may display, aggregate, merge, analyze, prepare material related to technology reports, ratings, rankings, technology trend data, or other data related to company technology, development, and/or engineering; Can be configured to report and distribute. CTO digital twin 8310 is capable of linking to, interacting with, and associating with external data sources, and uploading, downloading, and downloading external data sources, including internal data of the EMP, as described herein. You can aggregate and analyze this data. Data analytics, machine learning, AI processing, and other analytics may be coordinated between CTO digital twin 8310 and the analytics team based at least in part on using intelligent services system 8010. This collaboration and interaction may be seeded within the enterprise data repository 8012 for use in modeling, machine learning, and AI processing to identify an optimal technology strategy, or some other technology, development, and/or engineering-related metric or aspect. In addition to assisting with technology, development, and/or engineering-related data elements and domains, it may also include the identification of optimal data measurement parameters upon which to base judgment of the success of technology initiatives, development initiatives, and/or engineering efforts. You can. Examples of data sources 8020 that can be connected to, associated with, and/or accessed from the CTO digital twin 8310 include sensor systems 8022, sales databases 8024 that are updated with sales figures in real time, technology, A development, and/or engineering platform, a news website (8048), a technical database that tracks the costs of the business, an organizational chart (8034), a workflow management system (8036), a customer database (8040) that stores customer data, and/ or a third party data source 8038 that stores third party data.

In embodiments, CTO digital twin 8310 may aggregate data sources and types to create new data types, summaries, and reports that are not available elsewhere. This can reduce dependency on multiple third-party providers and the need for current solutions. This, among other benefits and improvements, can reduce the costs associated with obtaining the data needed for sound technical decisions.

In an embodiment, CTO digital twin 8310 includes real-time monitoring, based at least in part on the use of a monitoring agent in client application 8052, as described herein, associated with CTO digital twin 8310. , can be configured to monitor technical performance. The monitoring agent may report on these activities to EMP 8000 for presentation in a user interface associated with CTO digital twin 8310. In response, EMP 8000 trains an enforcement agent (which may include one or more machine learned models) to handle and process these notifications the next time they arrive, and determines whether these notifications are of an urgent nature, e.g. For example, the CTO may be escalated and/or alerted upon the identification of a new technical bug or security patch that is urgently needed. In embodiments, CTO digital twin 8310 may generate technical performance alerts based on performance trends. This can allow CTOs to optimize initiatives in real time without having to manually request such real-time technical performance data; CTO digital twin 8310 can automatically present this information and relevant/necessary alerts as configured by the organization, CTO, or some other interested party.

In embodiments, CTO digital twin 8310 may be used to determine performance of a technology, development, and/or engineering department, personnel, technology, development, and/or engineering activities, technology, development, and/or engineering departments, and /or may be configured to report on engineering content, technology, development, and/or engineering platforms, technology, development, and/or engineering partners, or some other aspect of management within the responsibility of the CTO. Reports may be to the CEO, technology, development, and/or engineering departments, other executives in the organization (e.g., CIO), or external third parties.

In embodiments, CTO digital twin 8310 may be configured to monitor, store, aggregate, merge, analyze, prepare, report, and distribute material related to industry best practices, benchmarks, or some other requirements or standards. For example, the CTO digital twin (8310) can communicate with other enterprise digital twins, such as the CIO digital twin (8312), which allows the technology team to keep the CIO informed when changes occur.

In embodiments, a client application 8052 executing a CTO digital twin 8310 enforces the CTO's behaviors and preferences (or the behaviors and preferences of other technical, development, and/or engineering personnel) as described herein. The execution agent system 8008 may consist of an executive agent reporting to an agent system 8008, which may be configured to assist the CTO or other technical, development, and/or engineering personnel in data collection, analysis, and machine learning as described herein. and A.I. Based at least in part on technology, executive agents can be trained on how to respond to specific situations and adjust their behavior.

In this example of CTO Digital Twin 8310, references to the features and functionality of the EMP and Digital Twin are to other departments and Digital Twins, and their respective projects and workflows, unless the context indicates otherwise. It must be understood as applicable.

In embodiments, the Chief Information Officer (CIO) digital twin 8312 is a digital twin configured for an enterprise's CIO, or similar executive responsible for overseeing the enterprise's intelligence, information, data, knowledge, and/or IT operations. You can. In an embodiment, CIO digital twin 8312 displays a real-time representation of an organization's information assets and workflows, including data related to data security, network security, and enterprise knowledge. A real-time representation may include at least real-time behavioral data that tracks the performance of an organization's information infrastructure, including internal information assets, customer-facing technologies, and information assets provided and/or serviced by third parties, such as cloud computing service providers. It can be partially based on For example, the CIO digital twin 8312 receives real-time information about the performance of the network, such as the intranet used by the organization, APIs accessed by the enterprise, APIs exposed by the enterprise, software running on enterprise software, etc. can do. The information can be aggregated and presented to the CIO to provide him with an overview of the general performance of the enterprise's computing infrastructure. For example, a CIO digital twin can indicate whether any network outages are occurring, whether there are any security risks detected in the corporate network, whether any software systems are behaving improperly, and other scenarios. In embodiments, CIO digital twin 8312 allows a user (e.g., a CIO) to select specific network assets to review in more detail, such as assets for which real-time operational data indicates that they are experiencing operational failures or other issues. A user interface can be presented. This real-time behavioral data related to the performance of IT and other information assets allows CIOs to better track the performance and needs of their organization's information and IT infrastructure, troubleshoot problems, simulate solutions, and take appropriate information and IT management actions. and maintain the organization's information and IT infrastructure.

In embodiments, CIO digital twin 8312 may provide real-time, historical, aggregated, comparative, and/or projected information (e.g., real-time, historical, simulated, and/or information related to company information and IT assets, third-party assets, etc.). or expected performance data), data, analysis, summaries, and/or information and IT reporting. CIOs using CIO Digital Twin (8312) can better maintain and evolve their information and IT assets through continuous monitoring using CIO Digital Twin (8312). CIO Digital Twin 8312 can assist with virtual monitoring and testing in a virtual environment to test implementation, changes, reconfiguration, introduction and/or removal of components and other assets, etc. In embodiments, the CIO digital twin operates in conjunction with EMP 8000 to perform simulation, prediction, statistical summarization, analysis, machine learning, and/or learning from other AI and inputs (e.g., performance data, sensor data, etc.). -Can provide decision support based on type processing.

In embodiments, the types of data that can populate CIO digital twin 8312 include information and IT asset performance and specification data, interoperability and compatibility data, cybersecurity data, uptime/downtime/operational performance data, asset aging/ It may include, but is not limited to, vintage/timing data, resource utilization, results of AI and/or machine learning modeling (e.g., IT performance simulation), or some other type of data related to the CIO's actions.

In embodiments, CIO digital twin 8312 may be configured to interface with collaboration suite 8006 to specify and provide a set of collaboration tools that technology, development, and/or engineering departments and associated parties can utilize. Collaboration tools may include video conferencing tools, “in-twin” collaboration tools, whiteboard tools, presentation tools, word processing tools, spreadsheet tools, etc., as described herein. Collaboration and communication rules may be constructed based at least in part on using an AI reporting tool, as described herein. Collaboration and communication tools and associated rules can be configured to use company-specific, industry-specific, and domain-specific taxonomies and vocabularies when representing entities, states, and flows within the CIO digital twin 8312.

In embodiments, CIO digital twin 8312 utilizes a network simulated under multiple virtual IT environments and scenarios that may impact performance, such as security breaches, IT asset failures, information failures, network congestion, or other activities or events. , can be configured to provide simulations of organizational information and IT activities, including but not limited to disaster planning, IT asset selection, maintenance protocols, downtime planning, etc. As described herein, real-time operational data, such as available through EMP, may be included in simulated information or IT infrastructure scenarios for the purpose of performing operational simulations. Simulation, analysis and/or modeling performed by EMP 100 on CIO digital twin 8312 may be used to reduce test time, design time, or some other type of IT cost. Simulation, analysis, and/or modeling performed by the CIO digital twin (8312) can be used to create and structure IT assets, networks, and guide development and test plans. Simulations, analysis and/or modeling performed by CIO digital twin 8312 may be used to evaluate network security, performance, and other characteristics. Equipped with a digital twin (8312), CIOs can quickly identify optimal asset configurations to maximize operational performance.

In embodiments, CIO digital twin 8312 may be configured to provide a user (e.g., a CIO) with information unique to CIO digital twin 8312, thereby providing real-time information based on both real-world and simulated activities. It may be based at least in part on the CIO Digital Twin 8312 using production, development and operational data to provide the CIO Digital Twin 8312 with unique information and insight and perspective on IT asset performance. In embodiments, CIO digital twin 8312 may be configured to manage operational plans based at least in part by utilizing predictive analytics for development plans. In embodiments, CIO digital twin 8312 may be configured to store, aggregate, merge, analyze, prepare, report, and distribute information and/or material related to IT strategies, scenarios, events, plans, activities, or initiatives. For example, CIO digital twin 8312 may provide a plurality of information, summaries, and reports and analyzes, including such materials, summaries, and reports and analyzes related to previous events, activities, and outcomes (e.g., system outages). It may be associated with a database or other repository.

In embodiments, CIO digital twin 8312 stores information and/or materials related to IT reporting, ratings, rankings, information, knowledge and IT trend data, or other data related to company information and/or IT assets and infrastructure. , can be configured to aggregate, merge, analyze, prepare, report, and distribute. The CIO Digital Twin 8312 is capable of linking to, interacting with and interacting with external data sources such that the CIO Digital Twin 8312 can upload, download, aggregate and/or analyze such enterprise data. can be related to

In embodiments, CIO digital twin 8312 may monitor IT data, including real-time, based at least in part on the use of a monitoring agent of a client application 8052, as described herein, associated with CIO digital twin 8312. Can be configured to monitor performance. The monitoring agent may report on these activities to EMP 8000 for presentation in a user interface associated with CIO digital twin 8312. In response, EMP 8000 (which may include one or more machine learning models) handles and processes these notifications the next time they arrive and escalates and/or alerts the CIO when these notifications are urgent. Enforcement agents can be trained.

In embodiments, CIO digital twin 8312 may be configured to report on the performance of an organization's IT assets, network, or some other aspect of management within the CIO's responsibilities. In an embodiment, a client application 8052 running a CIO digital twin 8312 may be configured with an executive agent that reports the CIO's behavior and preferences to an executive agent system 8008, which in turn or other personnel may perform the data collection, analysis, machine learning, and A.I. described throughout this disclosure. Based at least in part on technology, enforcement agents can be trained on how to respond to specific IT situations and adjust their behavior.

In this example of Marketing Department and CIO Digital Twin 8312, references to the features and functionality of the EMP and Digital Twin refer to other departments and Digital Twins and their respective projects and workflows, unless the context indicates otherwise. It must be understood as applicable.

In embodiments, general counsel (GC) digital twin 8314 may be an executive digital twin configured for an enterprise's general counsel (GC), or similar executive responsible for overseeing the enterprise's legal department and/or external counsel. . GC Digital Twin (8314) provides management of legal staff, partners and external counsel, oversight of legal budgets and resources, compliance, management of contracts and litigation, internal policies, intellectual property, employment law, tax law, privacy law, reporting, and regulation. May provide functionality including, but not limited to, management of analytics.

In embodiments, the types of data that may be populated and/or utilized by GC digital twin 8314 may include, but are not limited to: budget data (e.g., external legal spend, internal legal spend) , ancillary legal costs, etc.), regulatory data (e.g. regulatory requirements, regulatory actions taken, etc.); Contractual and licensing data (e.g. ongoing negotiations, current contractual obligations, past contractual obligations, etc.); Compliance data (e.g., compliance requirements, compliance actions taken, etc.), litigation data (e.g., potential sources of litigation, pending litigation, past litigation, settlement agreements, etc.), employment data (e.g., employment contracts, employee complaints, employee stock options, etc.), intellectual property data (e.g., filed patent applications, patent dockets, granted patents, trademark applications, trademark docket data, registered trademarks, etc.), tax data, privacy data, regulatory data, Analytical results from AI and/or machine learning modeling; predictive data; Referral data, or some other type of data related to the operations of GC and/or the Legal Department.

In embodiments, GC digital twin 8314 may be configured based, at least in part, on using collaboration suite 8006 to specify and provide a set of collaboration tools that the legal department and associated parties can utilize. Collaboration tools may include video conferencing tools, “in-twin” collaboration tools, whiteboard tools, presentation tools, word processing tools, spreadsheet tools, etc., as described herein. Collaboration and communication rules may be constructed based at least in part on using an AI reporting tool, as described herein. Collaboration and communication tools and associated rules can be used to represent entities, states, and flows within the GC digital twin 8314 using company-, industry-, and domain-specific taxonomies and vocabularies, such as corporate law, commercial law, bankruptcy law, securities transaction law, banking law, Customs law, export control regulations, maritime law, trade law, international treaties, securities law, contract law, environmental law, international law, privacy law, data privacy law, patent law, civil and criminal procedure, trademark law, copyright law, trade secret law, unfair competition law, tort law, and property law. , advertising laws, etc. may be configured for use in connection with a particular subject matter of a law, regulation, jurisdiction, or area of practice.

In embodiments, GC digital twin 8314 may support an overall department budget, a budget for a specific project, such as “US Patent Application,” or a group of projects, a budget for a specific litigation, or a budget for a third-party vendor, such as an external counsel. , or may be configured to research, generate, track, and issue reports on legal department budgets, including but not limited to, or some other type of legal budget. GC Digital Twin (8314) generates financial data related to materials under review or supervision of the Legal Department, including but not limited to licensing revenue, licensing expenditures, or some other type of financial data related to Legal Department review and accountability; It can be configured to track, deliver, research and report. In embodiments, GC digital twin 8314 may be used in conjunction with other executive twins as described herein, including but not limited to CFO digital twin 8304, CEO digital twin, COO digital twin, CTO digital twin, etc., and such licensing. Interact with and share revenue and/or budget data and reporting. In embodiments, GC digital twin 8314 can be used for data analytics, machine learning, and A.I., as described herein. Based at least in part on the process, for example, key departments, employees, third parties, or It may include intelligence that reads legal contracts, licenses, budgets and related summaries and data to identify others. License and/or budget material for a given party can be abstracted and summarized for presentation independently from the budget's totality, and formatted and presented automatically or at the user's direction to the party that is the subject of the budget item. In a simplified example, a GC may have license(s) with line items, schedules, appendices, etc. detailing the licensing revenue to be paid to the organization over a defined time frame, under review by its department. GC uses the GC Digital Twin (8314) to license revenue to other executives in the organization, such as the CFO (e.g., via a CFO digital twin) and/or the CEO (e.g., via a CEO digital twin). Such financial data derived from or to be derived from may be consolidated, summarized and/or shared. The shared data may indicate licensing revenue to be earned in a given financial quarter to help CFOs and others maintain an accurate and current summary of expected quarterly revenue.

In embodiments, GC digital twin 8314 may be configured to track and report inbound (e.g., settlement or litigation proceeds) and outbound claims (e.g., external counseling fees) associated with the legal department. Billing departments, employees, processes, and systems interact with the GC Digital Twin (8314) to present, store, analyze, and coordinate billing activity related to parties with whom the legal department is contracting, such as external counselors, consultants, research services, online entities, etc. and/or report. In embodiments, GC digital twin 8314 may be configured to research, track, monitor, store, analyze, create, and distribute legal content, and automatically report on such activities to a user interface associated with GC digital twin 8314. You can. Such activity may occur when the GC digital twin 8314 receives a status change, for example, a new court filing in a lawsuit, a communication received from an external counsel, a new license draft from an opposing counsel, a draft patent application, a notice from the U.S. Patent and Trademark Office, or It may involve storing data so that new or updated material of some other type can be detected. GC digital twin 8314 may also detect activity among classes of entities that are monitored or specified for monitoring in GC digital twin 8314, such as certain courts, regulatory or legislative bodies, or some other type of entity. . In embodiments, GC digital twin 8314 may be configured to research, track, monitor, store, and analyze content from various law-related platforms and automatically report on such activities to a user interface associated with GC digital twin 8314. You can. These platforms may include, but are not limited to, bar or other law associations, courts, legal search platforms, social media, legal blogs, press releases, or some other type of legal platform-related material or activity.

In embodiments, GC digital twin 8314 may be configured to store, aggregate, merge, analyze, prepare, report, and distribute material related to legal strategies, legal documents, lawsuits, legal recommendations, or some other legal activity. For example, GC digital twin 8314 may be associated with multiple databases or other repositories of legal materials, contracts, licenses, intellectual property (e.g., patent applications), summaries and reports and analyses. GC digital twin 8314 is capable of linking to, interacting with, and associating with external data sources, and uploading, downloading, and executing external data sources, including internal data of the EMP, as described herein. You can aggregate and analyze this data. Data analytics, machine learning, AI processing, and other analytics may be coordinated between GC digital twin 8314 and the analytics team based at least in part on using intelligent services system 8010. This collaboration and interaction enables enterprise data repositories (e.g., corporate data repositories) for use in modeling, machine learning, and AI processing to identify optimal and/or relevant legal content, legal documents, and parties involved in legal activities (e.g., litigation). 8012), as well as optimal data measurement parameters on which to base judgments on the success of legal efforts (e.g., licensing revenue, budget specified for use of external counsel). It may include identifying (residing within, etc.) Examples of data sources 8020 that can be connected to, associated with, and/or accessed from GC digital twin 8314 include legal research platforms, legal websites, news websites 8048, financial databases 8030, This may include, but is not limited to, a contract database, HR database 8046, workflow management system 8036, and/or a third party data source 8038 that stores third party data.

In embodiments, GC digital twin 8314 may be configured to assist in the development of new legal endeavors, such as pursuing new contracts, reviewing new laws or regulations affecting business, litigation, or arbitration, or some other legal activity. there is. For example, GC digital twin 8314 can identify teams of internal and external partners (e.g., external counsel) for legal action. For example, individuals who are ideal candidates to assist in legal action may be identified based at least in part on experience and expertise data stored within or in association with GC digital twin 8314. For example, a GC may initiate negotiations of a joint development agreement between an entity located in the United States and Taiwan and may need to seek counsel outside of Taiwan. Using the GC Digital Twin (8314), GC can be presented with details of previous external consultations used in Taiwan for similar projects. In another example, if GC Digital Twin 8314 cannot find details of previous external counselors used in Taiwan for similar projects, GC Digital Twin 8314 will work on a jointly developed agreed project based on skills, experience, etc. Scan, research, collect and summarize information from public or other sources about highly rated, recommended or other external Taiwan counselors as may be appropriate.

In an embodiment, GC digital twin 8314 identifies legal project objectives, records, monitors, and tracks the project's performance against those objectives, and reports the project to GC within a user interface associated with GC digital twin 8314. Tracking can be presented in real time. For example, GC digital twin 8314 may include a clickable dashboard that, when clicked, illustrates the status of a set of legal projects. In some embodiments, the dashboard may include a timeline for each project and each project's relative status with respect to the timeline.

In embodiments, GC digital twin 8314 may be configured to report on legal department performance, legal department employees, legal actions, legal content, legal platforms, legal partners, or some other aspect of GC's management. Reports may be to the GC, the legal department, another official of the organization (e.g., CEO), or an external third party (e.g., external counsel, legal notice, press release, etc.). Reports and their contents can be shared by GC Digital Twin 8314 with other enforcement digital twins, for example, data related to regulatory compliance, ongoing litigation, or some other legal activity. The reporting capabilities of GC Digital Twin (8314) can also be used to populate the necessary data for formal reporting requirements such as shareholder statements, annual reports, SEC filings, etc. Templates of common reporting formats can be stored and associated with the GC digital twin (8314) to automate the presentation of data and analysis according to predefined formats, styles and system requirements. In some embodiments, the GC digital twin may be configured to utilize trained enforcement agents 8364 on behalf of the GC to generate and disseminate reports.

In embodiments, GC Digital Twin 8314 monitors, stores, aggregates, merges, analyzes, and prepares material related to regulatory activities, such as government regulations, compliance, legislation, court opinions, industry best practices, or some other requirements or standards. , reporting and distribution. For example, the GC digital twin 8314 can keep informed of new regulations or regulatory changes as GC occurs. The GC may set parameters of the GC digital twin 8314 regarding legal domains, subject areas, jurisdictions, or some other parameters of interest to the GC that the GC digital twin 8314 should monitor.

In embodiments, GC digital twin 8314 may utilize an enforcement agent 8364 trained on the behaviors and preferences of users (e.g., GCs) (or the behaviors and preferences of other legal personnel). In embodiments, the client application 8052 hosting the GC digital twin 8314 may track the user's actions related to various events, notifications, alerts, etc., and the expert agent system 8008, as described herein. You can report tracked events using . In response, expert agent system 8008 can learn how a GC or other legal personnel respond to a particular situation, so that enforcement agent 8364 can respond to similar situations once deployed, allowing users (e.g. , an executive agent 8364 can be trained on behalf of GC).

In this example of the Legal Department and GC Digital Twin (8314), references to the features and functionality of the EMP and Digital Twin refer to other departments and Digital Twins, and their respective projects and It should be understood as applying to the workflow.

In embodiments, Chief Human Resources Officer (CHRO) Digital Twin 8316 (or HR Digital Twin 8316) is an enterprise human resources executive (e.g., CHRO) or It is an executive digital twin created for similar executives, such as Chief People Officers (CPOs), Chief Talent Managers, Human Resources Directors, and Human Resources Directors. In embodiments, CHRO digital twin 8316 may support human capital management, workforce management, risk management, and payroll, recruiting, compliance, employee performance, benefits, employee relations, time and attendance, training and development, compensation, and on-boarding. , can display different HR-related statuses of the company, such as status related to management of offboarding, succession planning, etc. In embodiments, CHRO digital twin 8316 may initially display various states at a lower level of granularity. Users viewing the CHRO digital twin (8316) can select which states to drill down to and view the selected states at a higher level of granularity.

In an example, Types of data that can be displayed in CHRO Digital Twin (8316) may include: This is not limited to: individual employee data; Key performance indicators by business unit; Key performance indicators for each individual employee; risk management data, Compliance data (e.g. OSHA and EPA compliance data); safety data, diversity data, Benefits data (e.g. Medical treatment, Dentist, vision, and health savings accounts (HSA) compensation data; compensation comparison data, compensation trend data, payroll data, overtime data, recruitment data, employee referral data; applicant data, Applicant screening data; Applicant reference data; Applicant background check data; offer data, Time and attendance data; employee relations data, employee complaint data; on-boarding data, offboarding data, staff training and development data; employee turnover data; voluntary employee turnover data; new hire turnover data; high performer turnover data; Turnover rate by performance rating data; Headcount and/or headcount planning data (e.g. number of people to plan percentage); promotion rate data, succession planning data, organization level data, control span data, employee survey data; Cost data for moving employees below the midpoint; comparison ratio data, simulation data, decision support data from AI and/or machine learning systems; Predictive data from AI and/or machine learning systems; Classified data from AI and/or machine learning systems; Detection and/or identification data from AI and/or machine learning systems, etc.

In embodiments, CHRO digital twin 8316 may display data items with icons that indicate whether the data item is in a normal state, suboptimal state, critical state, or alert state. In embodiments, icons may be different colors, fonts, symbols, codes, etc. For example, CHRO digital twin 8316 may display high performer turnover data with an orange icon indicating that high performer turnover is at a critical level. Continuing the example, an HR executive could be enabled to report high performer turnover data to another executive, such as the CEO, through the CHRO digital twin (8316). In embodiments, CHRO digital twin 8316 may automatically highlight data items that are suboptimal, critical, or in an alert state.

In embodiments, CHRO digital twin 8316 may be configured to provide an “in-twin” collaboration suite with tools that can facilitate communication and collaboration between enterprise stakeholders. In embodiments, an “in-twin” collaboration tool may include an interface that allows users to forward and/or counter-upload data sets to other users associated with the enterprise. In embodiments, the interface may be configured to allow a user to send messages with a data set, create requests or assign tasks related to the data set, and/or schedule events associated with the data set. In embodiments, AI and/or machine learning may be utilized to suggest message content, suggest event scheduling, suggest requests or tasks, and/or suggest request or task assignees. For example, an HR executive could tip off a data set related to employee training to the GC with predictive text messages and calendar requests regarding employee training at times determined by AI and/or machine learning to attend meetings related to employee training. there is. In an embodiment, the “in twin” collaboration tool includes digital twin conferencing. In embodiments, the “in-twin” collaboration tool may include an “in-twin” messaging system and/or an “in-twin” video conferencing system to enable enterprise stakeholders to communicate. In embodiments, machine learning and/or AI systems may be utilized to automatically create and/or assign tasks from such communications. In an embodiment, the “in-twin” video conferencing system supports subchat. In embodiments, subchats may be created via “drag and drop” actions in the user interface. In embodiments, an “in-twin” video conferencing system may utilize machine learning and/or AI to make suggestions for optimizing lighting, audio, camera placement, etc. for the user. In embodiments, an “in twin” video conferencing system utilizes machine learning and/or AI to automatically disable a video feed upon detection of inappropriate activity in the video feed. In an embodiment, the “in-twin” collaboration suite includes an “in-twin” stakeholder approval system to collect approval for actions from other enterprise stakeholders. In embodiments, an “in-twin” collaboration tool may include an AI-driven transformation system configured to intelligently transform communications between enterprise stakeholders to achieve maximum understanding by users of the digital twin, and AI-driven transformation. The system is configured to translate from a first language to a second language (eg, English to a foreign language) and to translate terminology or jargon so that a user can understand it. These features, described in relation to CHRO digital twin 8316, can be used for different types of enterprise control tower activities, such as monitoring operations, development activities, or other aspects of the enterprise across locations, departments, and functions. It can be deployed with other types of digital twins described in this application, including for other executives, including to facilitate collaboration between executives. Collaboration and communication tools and associated rules may include those related to worker health and safety, those related to education and training, those related to performance indicators, those related to worker attributes (including psychological, demographic and similar factors), etc. Likewise, the CHRO digital twin 8316 may be configured to use company-, industry-, and domain-specific taxonomies and vocabularies when representing entities, states, and flows.

In embodiments, CHRO digital twin 8316 may be configured to identify, interview, select, hire, and onboard new employees. In some of these embodiments, CHRO digital twin 8316 may include, but is not limited to, employee referral data, applicant training data, applicant testing data, applicant experience data, applicant reference data, applicant screening data, applicant background check data, It may be configured to research, track, and report applicant data, including applicant interview data, job application data, applicant resume data, applicant cover letters, applicant proposal data, and the like. CHRO digital twin 8316 may interact and share such applicant data and reporting with other executive digital twins, as described herein. CHRO Digital Twin (8316) uses job applications, resumes, cover letters, and applicant references to identify and select potential new employees and/or identify other executives or corporate stakeholders who may be interested in such information. , may include other intelligence such as machine learning, AI, and/or analytics to process applicant screening data, applicant interview data, etc.

In embodiments, EMP 8000 (e.g., via an API) may be used to support an enterprise's human resources management software, human capital software, workforce management software, payroll software, applicant tracking software, accounting software, employee applicant software, publicly disclosed financial services, etc. HR-related data can be obtained from statements, third-party reporting, tax filings, social media software, job listing websites, recruiting software, and more.

In embodiments, CHRO digital twin 8316 may provide an interface for HR executives to perform one or more HR-related workflows. For example, CHRO digital twin 8316 can be used in the on-boarding workflow, offboarding workflow, termination workflow, decision document workflow, succession planning workflow, candidate evaluation workflow, candidate screening workflow, and compliance workflow. HR-executives to perform, supervise, or monitor workflows, entities accompanying workflows, and their attributes, such as disciplinary workflows, review workflows, interview workflows, proposal workflows, employee training workflows, etc. An interface for can be provided.

In an embodiment, CHRO digital twin 8316 uses expert agent system 8008 to respond to events and situations encountered by users (e.g., alerts, notifications, e-mails, delegations, presentation of data, events, etc.) An executive agent 8364 may be utilized that is trained on the actions (e.g., behaviors, responses, interactions, and preferences) of users (e.g., HR executives). In some of these embodiments, the client application 8052 hosting the CHRO digital twin 8316 may report actions taken by the user in response to various events the user encounters through the CHRO digital twin 8316. For example, client application 8052 may identify events such as a request to authorize a new hire, a request to terminate an employee, or a notification indicating that employee turnover has reached a critical threshold. In this example, the client application 8052 may record and report actions taken by the user in response to such events, actions associated with the identified event, as well as any other characteristics associated with the event to the expert agent system. 8008). In response, expert agent system 8008 can train executive agent 8364 on behalf of the user so that the executive agent can perform actions or make recommendations to the user when encountering similar events in the future.

References to the features and functionality of the EMP and Digital Twin in this example of Human Resources and CHRO Digital Twin (8316) refer to other departments and Digital Twins, and their respective projects and It should be understood as applying to the workflow.

In embodiments, an executive digital twin may connect to, interact with, integrate with, and/or be used by multiple different applications. For example, the executive digital twin may be used in an automated AI-reporting tool (8360), in a collaboration tool (8362), in relation to an executive agent (8364), in a board meeting tool (8366), in a training module (8368), And may be used for planning tool 8370.

In embodiments, AI reporting tool 8360 assists users in reporting one or more conditions to other users. For example, subordinates may need to report identified problems to higher-up members of the company (for example, the CTO may want to report problems that need to be addressed to the CEO). In an embodiment, AI reporting tool 8360 may be configured to receive a request to report a status from client device 8050. In embodiments, AI-reporting tool 8360 may identify appropriate recipients of the reported status based on the type of request, the role of the user issuing the request, and the organizational structure of the entity. In some embodiments, AI-reporting tools can determine the roles of users and recipients of reports from an enterprise's organizational digital twin. In some embodiments, AI-reporting tool 8360 may determine whether the intended recipient of the notification has access to the data being shared from the executive digital twin. For example, if the CFO reports to the CEO, the CEO will likely have access to all company data and will not be excluded from receiving reports. Conversely, if the CFO wants to delegate handling of issues through an AI-reporting tool to employees of his or her business unit, the recipient may not have access to this data. In this scenario, AI-reporting tool 8360 can notify the requesting user (e.g., CFO) that certain types of data may not be shared with junior employees and resolve the issue without sharing inaccessible data. You can decide how lower-level employees will be reported. Once it determines that the user has access to view a particular state of data, AI-reporting tool 8360 can generate a report that is specific to the intended recipient. In embodiments, the AI-reporting tool may utilize the intelligent system's NLP services to generate reports. In some embodiments, AI-reporting tool 8360 may utilize enforcement agent 8364 to determine when to report a condition and the appropriate recipient of the reported condition. In this embodiment, enforcement agent 8364 may be trained on the user's interactions with the digital twin and client application 8052 previously presented to the user.

In some embodiments, AI-reporting tool 8360 may be configured to monitor one or more user-defined key performance indicators (KPIs). Examples of KPIs for an enterprise may include, but are not limited to, the following in relation to a system, facility, process, function, or workforce unit: uptime (e.g., of an assembly line or other manufacturing system), capacity utilization; , standard operating efficiency, overall effectiveness, downtime, amount of unscheduled downtime, setup time, amount of inventory turns, inventory accuracy, quality metrics related to products and services, amount of first pass yield for the enterprise, and required material. Volume of operation, days-sales-outstanding (DSO), amount of scrap or waste produced, throughput, conversion, maintenance percentage, yield per system or unit, overall yield, industry reviews, industry ratings, customer reviews, customer ratings, Editorial reviews, awards, social media and website attention metrics, search engine performance metrics, safety metrics, health metrics, environmental impact metrics, political metrics, certification and testing metrics, regulatory metrics, social impact metrics, financial and investment metrics, corporate bond ratings. , trade association metrics, coalition metrics, lobbying organization evaluation, advertising performance metrics, referral metrics, etc. Additional or alternative KPI metrics may be defined by the user. Examples of these KPI metrics include volume or percentage of failed audits, number or percentage of on-time/late deliveries, number of customer returns, number of employee training hours, percentage of employee turnover, number of reportable health or safety incidents, revenue per employee, and number of employees per employee. May include profit, schedule completion metrics, total cycle time, etc.

In embodiments, collaboration tools 8362 include various tools that allow for collaboration between executives of an enterprise. In an embodiment, the collaboration tool includes digital-twin enabled video conferencing. In this embodiment, EMP 8000 may present a requested view of the enterprise digital twin to participants in a video conference. For example, during a board meeting, a CTO proposing updates to machinery or equipment within a facility could present a digital twin of the facility's environment where the updates to the machinery or equipment would be made. In this example, the CTO can illustrate the results of simulations performed on the facility without and with updates. The simulation can illustrate how an update could benefit the enterprise using a number of selected metrics (e.g., throughput, profit, employee safety, etc.). Collaboration and communication tools and associated rules can be configured to use company-specific, industry-specific, and domain-specific taxonomies and vocabularies when representing entities, states, and flows within a digital twin.

In an embodiment, executive agent 8364 is an expert agent that is trained to perform tasks on behalf of executive users. As discussed, in some embodiments, a client application may monitor a user of the client application by the user when using the client application 8052. In this embodiment, the client application 8052 may monitor the status of the executive digital twin that the user drills down into, the status that the user reports to superiors and/or delegates to team members within each business unit, decisions being made, etc. . As a user uses the client application 8052, the expert agent system 8008 can train one or more machine learned models on behalf of that particular user, such that the model can perform tasks on behalf of the user or It may be utilized by the executive agent 8364 to recommend actions.

In embodiments, board meeting tools 8366 can be used to prepare for, access and access board and similar meetings, such as board meetings, boards of trustees, shareholder meetings, annual meetings, investor meetings, and other important meetings. / or a tool used to follow up on it. References to board meetings in this application should be understood to include those and other important meetings requiring executive preparation, attendance and/or attention. In embodiments, board meeting tool 8366 may allow different users to present one or more statuses of the enterprise digital twin within the context of a board report or board meeting. For example, a user (e.g., a COO) may view a simulation of a proposed logistics solution from the COO digital twin 8366 on one or more devices (e.g., a device within a board room and/or remotely accessing a board meeting). can be shared with the participant's device). In embodiments, board meeting tools 8366 may restrict access to certain types of data based on time, scope, and permissions. For example, board meeting tool 8366 may allow some or all of the data displayed in the digital twin being presented only on devices in one of the registered geolocations and/or only for a defined duration, such as before the meeting. Any geographic location of board members prior to a board meeting (e.g., board room, designated home office for those joining via phone or video, etc.), so that they can be visible from a few hours to a few hours after the meeting, or only during the meeting. It may be necessary to register. Similarly, in embodiments, board meeting tool 8366 may limit access to some or all of the data shared in a presented digital twin to certain times (e.g., during a board meeting or on the day of the board meeting). Other examples of board meeting tools 8366 are discussed throughout this application.

In embodiments, training module 8368 may include software tools used to train users. In embodiments, training module 8368 may utilize digital twins to improve executive training for enterprises. For example, training module 8368 may provide real-world examples based on data collected from businesses. Training module 8368 can present different scenarios to the user through executive digital twin 8368, and the user can take action. Based on the action, training module 8368 can request a simulation from EMP 8000, which in turn returns results to the user. In this way, the user can be trained for scenarios based on the user's actual enterprise.

In an embodiment, planning tool 8370 is a software tool that utilizes digital twins to help users plan for their enterprises. In embodiments, planning tool 8370 may be configured to provide a graphical user interface that allows executives to perform planning (e.g., defining budgets, KPIs, etc.). In some embodiments, planning tool 8370 may be configured to request a simulation from IMP 8000 given parameters set in the generated plan. In response, EMP 8000 may return the results of the simulation and the user may decide whether to adjust the plan. In this way, the user can iteratively refine the plan to achieve one or more goals. In embodiments, the executive agent 8362 may monitor tracking actions taken while a plan is being refined by a user, such that the expert agent system 8008 may direct the executive agent to create or recommend a plan to the user in the future. (8362) can be trained.

The enterprise digital twin may interface with and/or be utilized in conjunction with other software applications without departing from the scope of this disclosure.

84 illustrates an example implementation of EMP 8000. In this example, EMP 8000 communicates with a plurality of client applications 8052 and a set of enterprise assets 8400. In an example, EMP 8000 may be associated with a sensor system 8022, a physical entity 8402, a digital entity 8404, a computational entity 8406, and/or a network entity 8408 belonging to and/or associated with an enterprise. Receive corporate data from the same set of corporate entities (8400). In embodiments, enterprise data may relate to an enterprise's environment, processes, and/or conditions. For example, sensor system 8022 may be used at an enterprise's enterprise facilities (e.g., manufacturing facilities, warehouses, distribution centers, logistics facilities, transportation facilities, office buildings, customer locations, retail locations, agricultural facilities, natural resource extraction facilities). etc.), whereby sensor system 8022 may be configured to receive sensor readings (e.g., vibration data, location data, motion data) generally associated with a facility or a machine, piece of equipment, or other physical or workforce asset within the facility. , temperature data, pressure data, etc.). Within a facility, multiple physical assets (e.g., robots, autonomous vehicles, smart equipment, personnel, etc.) or other entities may output data streams related to the operation of the assets or other entities. Additionally or alternatively, an enterprise may include multiple digital assets (e.g., CRM, ERP, databases, etc.) that provide data streams related to sales, costs, human resources, etc. Network entities can provide networking-related data including bandwidth, API requests, throughput, detected cyberattacks, etc. Computational entities may provide data about an enterprise's computing infrastructure. In some embodiments, enterprise management system 8000 may also receive data from other sources, including third-party data 8038 from third-party data providers. Taken in combination, data from enterprise assets 8400 and/or other data sources may be used to determine the status of industrial facilities and machinery contained therein, various processes (e.g., industrial processes, sales workflows, hiring processes, logistics operations). It can provide information on the status of the business (e.g., flow, etc.), the efficiency of the process, and the financial health of the company.

In embodiments, corporate entities may communicate directly with EMP 8000 over a communications network. Additionally or alternatively, one or more of the enterprise assets may stream data to a local data collection system 8420 that collects and stores enterprise data locally. In some embodiments, local data collection system 8420 may provide collected data to an enterprise edge intelligence system 8422.

In embodiments, edge intelligence system 8422 may be a local data collection system 8420, a local sensor system 8022, or another enterprise entity located at or near the entity's physical location (e.g., at an industrial facility). Can be executed by an edge device 8042 configured to receive data, such as from 8400, and can perform one or more edge-related processes related to the received data. Edge devices may be pre-configured and/or substantially self- or automatically configuring computing devices, such as “edge intelligence in a box” devices. Edge-related processes may refer to processes performed on edge devices to store sensor data, reduce bandwidth on communication networks, and/or reduce computational resources required in backend systems. Examples of edge processes may include data filtering, signal filtering, data processing, compression, encoding, fast prediction, fast notification, emergency alerting, etc., and may include the creation of automated smart data bands. For example, edge intelligence system 8422 may determine whether to transmit a subset of data to EMP 8000 or store the subset of data locally until explicitly requested from EMP 8000. In another example, edge intelligence system 8422 may be configured to compress a data stream (e.g., a sensor data stream) to improve data throughput of a large data stream (e.g., vibration data). In some embodiments, edge intelligence system 8422 may be configured to analyze large amounts of data to determine whether to compress or stream the raw data stream. In some embodiments, local data collection system 8420 and edge intelligence system 8422 may be implemented on an enterprise edge device 8042. In some embodiments, edge intelligence system 8422 may communicate data to EMP 8000. In some of these embodiments, edge intelligence system 8422 communicates data to EMP 8000 via network enhancement system 8424.

In embodiments, network enhancement system 8424 may be configured to optimize the flow of data transmitted from and received by EMP 8000 from one or both edge intelligence system 8422 and local data collection system 8420. there is. For example, local data collection system 8420 may be configured to collect data from one or more real-world environments, entities, ecosystems, and/or processes that may be analyzed by connected edge intelligence system 8422. In this example, edge intelligence system 8422 may transmit the collected data to network enhancement system 8424, which may optimize transmission of the data to EMP 8000 for processing and implementation by EMP 8000. You can. The EMP 8000 allows the client application 8052 to update an enterprise digital twin (e.g., role-based digital twin, environment digital twin, cohort digital twin, etc.) hosted by the client application 8052. Data transmitted to 8052 may be stored, analyzed, or otherwise processed.

In embodiments, network enhancement system 8424 may include one or more signal amplifiers, signal headers, digital filters, analog filters, digital-to-analog converters, analog-to-digital converters, and/or antennas configured to optimize the flow of data. . In some embodiments, the network enhancement system may include a wireless repeater system such as that disclosed in U.S. Pat. No. 7,623,826 to Pergal, the entire contents of which are incorporated herein by reference. Network enhancement system 8424 may, for example, filter data, repeat data transmissions, amplify data transmissions, adjust one or more sampling rates and/or transmission rates, and implement one or more data communication protocols. The flow of data can be optimized.

In embodiments, network enhancement system 8424 may include one or more processors configured to perform digital signal processing to optimize the flow of data. One or more processors may implement optimization algorithms to optimize the flow of data. One or more processors may determine one or more optimal paths in the network, and network enhancement system 8424 transmits data along the one or more optimal paths. Network enhancement system 8424 may be configured to implement software filters through one or more processors. Software filters may filter data prior to transmission to EMP 8000, for example, to lower the network bandwidth consumed by data transmission. One or more processors determine that a portion of data is relevant only to one or more intended recipients, such as a digital twin, enforcement agent, collaboration suite, or other component of EMP 8000 and determine an optimal path based on the intended recipients of the portion of data. You can decide.

In embodiments, network enhancement system 8424 may be configured to optimize data flow between multiple nodes over multiple data paths. In some embodiments, network enhancement system 8424 may transmit a first portion of data over a first of the plurality of data paths and a second portion of data over a second of the plurality of data paths. Network enhancement system 8424 may determine that one or more data paths, such as a first data path, a second data path, or another data path, are favorable for transmission of one or more portions of data. Network enhancement system 8424 may determine one or more types of data being transmitted, one or more protocols suitable for transmission, current and/or expected network congestion, timing of data transmission, and current and/or expected volume of data being transmitted or to be transmitted. A favorable data path may be determined based on one or more networking variables, such as the like. Protocols suitable for transmission may include transmission control protocol (TCP), user datagram protocol (UDP), etc. In some embodiments, a network enhancement system may be configured to implement methods for data communication such as those disclosed in U.S. Pat. No. 9,979,664 to Ho et al., the entire contents of which are incorporated herein by reference.

EMP 8000 receives enterprise data (e.g., directly or through network enhancement system 8424, edge intelligence system 8422, local data collection system 8420, or from any other data source). In embodiments, digital twin system 8004 structures and/or stores enterprise data in one or more digital twin databases (e.g., graph databases, relational databases, SQL databases, distributed databases, blockchains, caches, servers, etc.). You can save it. In an embodiment, client application 8052 requests enterprise digital twin 8410 from EMP 8000. In response, the digital twin system 8004 generates the requested enterprise digital twin 8410 (e.g., role-based digital twin, executive digital twin, environmental digital twin, process digital twin, cohort digital twin, etc.) to the client. May serve application 8052, where enterprise digital twin 8410 may include enterprise data and/or data derived from enterprise data (e.g., by an intelligent service system). Client application 8052 may provide an interface for users of client application 8052 to interact with the requested digital twin 8410. For example, a user can delegate tasks related to the displayed status to subordinates and/or notify a superior of the displayed status via the digital twin interface. In another example, a user can drill down to a specific state and initiate corrective action through the digital twin interface. In some embodiments, client application 8052 allows users to share digital twin 8410 (or portions thereof) within collaboration tool 8414 or access collaboration features of collaboration tool 8414 within twin 8410. can be allowed. For example, client application 8052 may allow a user to share a displayed state of digital twin 8410 to a board meeting collaboration tool. Additionally or alternatively, expert agent 8364 may monitor the user's interactions with the digital twin and report the interactions to the EMP's expert agent system 8008. In an embodiment, expert agent system 8008 can receive interactions and train expert agent 8364 based on the interactions with the digital twin, as well as results resulting from the expert agent. For example, an expert agent could be trained to identify situations in which a user delegates a task or notifies a superior.

The executive digital twin discussed in connection with FIG. 71 is provided as an example and is not intended to limit the scope of the present disclosure. Additional and/or alternative data types may be included in each type of executive digital twin.

Figure 73 illustrates an example method 8510 for configuring and serving an enterprise digital twin. In embodiments, the method may be implemented by digital twin system 8004. The method can be performed on different types of enterprise digital twins, including role-based digital twins (e.g., executive digital twins), cohort digital twins, environmental digital twins, process digital twins, etc.

At 8512, a structural view for a specific type of digital twin is selected. In embodiments, structural views may be stored in a graph database (representing interconnected data) or a geospatial database (representing the coordinates of an actual facility).

At 8514, associated transaction data for the digital twin is selected. In embodiments, a combination of interaction data and transaction data is selected at a grain suitable for dynamic interaction within the digital twin. This selection process may involve dynamic configuration of the structure, functionality, and features of a data mart or other abstraction system and/or may operate dynamically, typically using a high-performance database storage mechanism (e.g., a columnar database or an in-memory database). there is.

At 8516, embellishment and/or augmentation data for the digital twin is selected. In embodiments, decoration data is an associated attribute that can be bound to an element within an executive digital twin. For example, in creating a digital twin of a facility's environment, decorative or augmented data can include the age of machinery or other assets within the facility, the names of key third-party suppliers whose items may be substituted for supply chain delivery, and It may include the input or output of the process flow that occurs, the identity of the manager, and indicators of the status and flow. In an abstract executive digital twin, embellishment data may include social media data, for example, sentiment analysis that may be associated with a customer hierarchical view.

At 8518, a representation medium for the digital twin is selected. In an embodiment, the final representation may be multifaceted, including a range of devices from simple mobile phone-based devices and touchscreen tablets to special purpose devices and/or immersive AR/VR headsets, among others. The presentation medium preferably influences the volume and nature of the data selected in the previous steps. In embodiments, selection of presentation medium is provided as a feedback indicator to the data and networking pipeline, so that filtering and data path selection can be performed recognizing the different capabilities and requirements of the end device and presentation medium. This may occur automatically, for example, by an agent being trained to provide context-sensitive feedback based on a training set of results.

At 8520, a perspective view is constructed. In an embodiment, perspective builder 8110 creates levels and characteristics of data that allow different types of users to interact with the digital twin while obtaining appropriate perspective levels. For example, from a CEO-level perspective, the CEO may request the context of third-party alternatives, market forces, and current strategic initiatives. In this example, viewpoint builder 8110 takes these considerations into account in creating the level of digital twin appropriate for the CEO, which will also impact the data selection process as different data grains are appropriate for different views. These different perspectives can be interacted with multiple roles simultaneously, allowing executives to view and interact with information relevant to their specific needs while providing guidance on the same topic.

At 8522, user notification is enabled. In embodiments, notifications within the digital twin are controlled by the grain of data selected and the viewpoint required. For example, the CTO level view requires notification of various technological changes and technological market forces, and the CTO digital twin is continuously overlaid with these notifications, which are structurally associated with relevant parts of the digital environment abstract or concrete. For example, in an organizational chart, a CTO can see implementation options for new technologies to provide more efficient communication between organizational units in a strategic planning exercise to acquire a new company. Simultaneously, CFOs are seeing the financial implications of these various options, and CEOs are being informed of decisions that could impact future market opportunities regarding upcoming company acquisitions.

This method is provided as an example only. Additional and/or alternative methods may be implemented to create and serve digital twins without departing from the scope of this disclosure.

The method of Figure 73 is provided as an example and is not intended to limit the scope of the present disclosure. Methods may include additional or alternative acts.

Figure 74 illustrates an example set of operations for a method 8600 for constructing an organizational digital twin. In embodiments, the method may be implemented at least in part by digital twin system 8004. It is understood that the methods may be implemented by other suitable computing systems without departing from the scope of the present disclosure.

At 8610, the organization chart of the company is determined. In embodiments, a user may upload an organizational chart through a GUI displayed to the user. In some embodiments, digital twin system 8004 or connected components may crawl one or more websites (e.g., corporate websites, social networking websites, etc.) and parse the crawled website(s) to You can decide on the organization chart.

At 8612, the enterprise's organizational framework is updated based on user input. In embodiments, a user may define roles within an enterprise to individuals listed in an organizational chart, may grant access to different roles and/or individuals, may grant permissions to individuals and/or roles, and may grant permissions to individuals and/or roles. and/or define relationships between individuals. In embodiments, relationships may represent reporting structures, teams, business units, etc.

In 8614, an enterprise's organizational digital twin is created and deployed. In embodiments, digital twin system 8004 may create an organizational digital twin by linking data from an enterprise to an organizational chart. This may include information about the individual, such as date of birth, social security or tax id, role, relationships, citizenship, employment status, salary, stock holdings, job title, current status, goals or targets, etc. Once deployed, the organizational chart can be continuously updated from one or more corporate data sources. In embodiments, an organizational digital twin may be utilized to determine the digital twin's reporting structure and/or the roles of individuals within the organization.

The method of Figure 74 is provided as an example and is not intended to limit the scope of the present disclosure. Methods may include additional or alternative acts.

Figure 75 illustrates an example set of operations of a method 8700 for creating an executive digital twin. In embodiments, the method may be implemented at least in part by digital twin system 8004. It is understood that the methods may be implemented by other suitable computing systems without departing from the scope of the present disclosure.

At 8710, a request for an executive digital twin is received from a user. In embodiments, digital twin system 8004 may receive a request for an executive digital twin from a user device associated with the user, such as a mobile device, personal computer, VR device, etc. The request may indicate the user's identity and/or the user's role.

At 8712, the user's role is determined. In embodiments, digital twin system 8004 may determine a user's role from a request and/or from an organizational digital twin of an enterprise associated with the user. In embodiments, an organizational digital twin may display the user's role, the user's permissions, the user's access rights, the user's restrictions, and the user's reporting structure.

At 8714, the configuration of the executive digital twin is determined based on the user's role. In embodiments, the configuration of the executive digital twin represents the granularity of the digital twin and the set of states to be represented in the executive digital twin. In embodiments, the configuration of the executive digital twin is stored in a configuration file within a digital twin data repository associated with the enterprise. The configuration file can define the initial state of the digital twin and the granularity of the state.

At 8716, a digital twin is created based on one or more data sources corresponding to the enterprise. In embodiments, digital twin system 8004 may determine an appropriate perspective for a requested digital twin based on the configuration of the digital twin and any access rights or restrictions of the user. In embodiments, restrictions may include limits on the data a user may have, limits on interaction, limits on the depth of data, limits on usage, and limits on the length of visibility. In some embodiments, generating a requested digital twin may include identifying appropriate data sources for the digital twin, given the perspective, and obtaining arbitrary data from the data sources to initially parameterize the executive digital twin. You can.

At 8718, the executive digital twin is served to the user's user device. In embodiments, digital twin system 8004 may provide a file (e.g., a JSON file) containing executive digital twin data and any data structures or visual elements necessary to display the executive digital twin by a user device. You can. In an embodiment, the digital twin system 8004 may also provide one or more real-time or near-real-time data streams (e.g., data via bus) can be streamed to the user device. Users can then interact with the digital twin. For example, users can delegate tasks through the executive digital twin, request simulations through the executive digital twin, drill down or zoom out on status displayed in the executive digital twin, and report status to a supervisor through the executive digital twin. You can do things like:

The method of Figure 75 is provided as an example and is not intended to limit the scope of the present disclosure. Methods may include additional or alternative acts.

Artificial Intelligence and Neural Network Embodiments

76-103, in embodiments of the present disclosure including those involving artificial intelligence 1160, expert systems, self-organization, machine learning, automation (robotic process automation, remote control, autonomous operation, automation) adaptive intelligence and adaptive intelligence systems, including prediction, classification, optimization, etc., for pattern recognition, classification of one or more parameters, characteristics, or phenomena, for support of autonomous control, and Other purposes can benefit from the use of neural networks or other artificial intelligence systems, such as trained neural networks. Throughout this disclosure, references to artificial intelligence, neural networks or neural nets include feedforward neural networks, radial basis function neural networks, self-organizing neural networks (e.g., Kohonen self-organizing neural networks), and recurrent neural networks. Neural networks, modular neural networks, artificial neural networks, physical neural networks, multilayer neural networks, convolutional neural networks, hybrids of neural networks with other expert systems (e.g. hybrid fuzzy logic-neural network systems), autoencoder neural networks, stochastic neural networks, time delay neural networks. , convolutional neural network, modulated feedback neural network, radial basis function neural network, recurrent neural network, Hopfield neural network, Boltzmann machine neural network, self-organizing map (SOM) neural network, learning vector quantization (LVQ) neural network, fully recurrent neural network, simple recurrent neural network, Echo state neural network, long short-term memory neural network, bidirectional neural network, hierarchical neural network, stochastic neural network, genetic scale RNN neural network, machine committee neural network, association neural network, physical neural network, momentary trained neural network, spiking neural network, neo-cognitive neural network, dynamic neural network, Cascading neural networks, neural-fuzzy neural networks, configural pattern generation neural networks, memory neural networks, hierarchical temporal memory neural networks, deep feedforward neural networks, gating recurrent unit (GCU) neural networks, autoencoder neural networks, variational autoencoder neural networks, noise-removing auto Encoder neural networks, sparse autoencoder neural networks, Markov chain neural networks, restricted Boltzmann machine neural networks, deep trust neural networks, deep convolutional neural networks, deconvolutional neural networks, deep convolutional inverse graphics neural networks, generative adversarial networks, liquid state machine neural networks, extreme learning machines. Neural networks, echo state neural networks, deep residual neural networks, support vector machine neural networks, neural Turing machine neural networks, and/or holographic associative memory neural networks, or hybrids or combinations of the foregoing, or rule-based systems, model-based systems (physical models understood to include a wide range of different types of neural networks, machine learning systems, artificial intelligence systems, etc., in combination with other expert systems, including those based on statistical models, flow-based models, biological models, biomimetic models, etc. shall.

The neural network described above may have various nodes or neurons, which may perform various functions on input, such as input received from sensors or other data sources, including other nodes. A function may involve weights, features, feature vectors, etc. Neurons may include perceptrons, neurons that mimic biological functions (such as the human senses of touch, sight, taste, hearing, and smell). Continuous neurons, such as sigmoidal activation, can be used in the context of various types of neural networks, such as those involving backpropagation.

In many embodiments, an expert system or neural network may be trained, such as by a human operator or supervisor, or based on a data set, model, etc. Training involves not only one or more training data sets representing values (including many of the types described throughout this disclosure), but also one or more sets of results, such as the result of a process, the result of a computation, the result of an event, the result of an activity, etc. It may involve presenting indicators to a neural network. Training a neural network to optimize one or more systems based on one or more optimization approaches such as Bayesian approach, parametric Bayes classifier approach, k-nearest-neighbor classifier approach, iterative approach, interpolation approach, Pareto optimization approach, algorithmic approach, etc. This may include training in optimization, such as training . Feedback may be provided in a process of variation and selection, such as using a genetic algorithm to evolve one or more solutions based on feedback over a series of rounds.

In embodiments, receiving data streams and other inputs collected from one or more environments (e.g., by a mobile data collector) and transmitted to a cloud platform over one or more networks, including using network coding to provide efficient transmission. Multiple neural networks may be deployed on a cloud platform. In the cloud platform, a plurality of different neural networks of various types (including modular form, structurally adaptive form, hybrid, etc.), optionally using massively parallel computing power, perform prediction, classification, and control functions, and the present disclosure It may be used to provide other outputs as described with respect to the expert system disclosed throughout. Different neural networks can be selected, for example by an expert system, for a particular task involved in a given situation, workflow, environmental process, system, etc., so that an appropriate type of neural network with appropriate input set, weights, node types and functions, etc. They can be structured to compete with each other (optionally including using evolutionary algorithms, genetic algorithms, etc.).

In embodiments, the methods and systems described herein involving expert systems or self-organizing capabilities may operate in one direction from a data input, such as a source of data about an individual, to an output, through a series of neurons or nodes. You can use a feedforward neural network that moves information to . Data can move from input nodes to output nodes, and optionally through one or more hidden nodes without looping. In embodiments, a feedforward neural network may be composed of various types of units, such as binary McCulloch-Pitts neurons, the simplest of which is a perceptron.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be used in some situations involving interpolation in a multidimensional space (e.g., when interpolation helps optimize a multidimensional function). RBF (which may be desirable, e.g., when optimizing data markets as described in this application, when optimizing the efficiency or output of power generation systems, factory systems, etc., or in other situations involving multiple dimensions) A radial basis function (radial basis function) neural network can be used. In an embodiment, each neuron within the RBF neural network stores examples from the training set as “prototypes.” The linearity involved in the functioning of this neural network gives RBF the advantage of not typically suffering from local minima or maxima problems.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may utilize radial basis function (RBF) neural networks, such as those using a distance to centroid criterion (e.g., a Gaussian function). there is. Radial basis functions can be applied as a replacement for hidden layers, such as sigmoidal hidden layer propagation in multilayer perceptrons. An RBF network may have two layers, such that inputs are mapped onto each RBF in a hidden layer. In an embodiment, the output layer may include, for example, a linear combination of hidden layer values representing the average predicted output. Output layer values may provide outputs that may be the same or similar to those of a regression model in statistics. In a classification problem, the output layer may be a sigmoid function of a linear combination of hidden layer values representing the posterior probability. In both cases, performance can often be improved by shrinkage techniques, such as ridge regression in classical statistics. This corresponds to the a priori confidence in small parameter values (and therefore smooth output functions) in the Bayesian framework. RBF networks can avoid local minima because the only parameter adjusted in the learning process is the linear mapping from the hidden layer to the output layer. Linearity ensures that the error surface can be quadratic and therefore has a single minimum. In regression problems, this can be found in one matrix operation. In classification problems, the fixed nonlinearity introduced by the sigmoid output function can be treated using iterative use. Reweighted least squares function, etc.

In embodiments, the RBF network may use kernel methods such as support vector machines (SVM) and Gaussian processes (where RBF may be a kernel function). A non-linear kernel function can be used to project the input data into a space where the learning problem can be solved using a linear model.

In an embodiment, an RBF neural network may include an input layer, a hidden layer, and a summation layer. In the input layer, one neuron appears in the input layer for each predictor. For categorical variables, N-1 neurons are used, where N is the number of categories. The input neuron may, in an embodiment, normalize the range of values by subtracting the median and dividing by the interquartile range. The input neuron can then feed the value to each neuron in the hidden layer. In the hidden layer, a variable number of neurons can be used (determined by the training process). Each neuron can be composed of a radial basis function that can be centered at a point with as many dimensions as the number of predictors. The spread (eg, radius) of the RBF function may be different for each dimension. Center and spread can be determined by training. When presented with a vector of input values from the input layer, the hidden neuron can compute the Euclidean distance of the test case from the neuron's center point and then apply the RBF kernel function to this distance, for example using the diffusion value. there is. The resulting value can then be passed to the summation layer. In the summation layer, the value originating from a neuron in the hidden layer can be multiplied by the weight associated with the neuron and added to the weighted values of other neurons. This sum becomes the output. For classification problems, one output (with a separate set of weights and summing units) can be generated for each target category. The value output for a category is the probability that the case being evaluated has that category. In the training of RBF, the number of neurons in the hidden layer, the coordinates of the centroid of each hidden layer function, the spread of each function in each dimension, and the weights applied to the output when it is passed to the summation layer. Various parameters can be determined. Training can be used by clustering algorithms (such as k-means clustering), evolutionary approaches, etc.

In embodiments, a recurrent neural network may have time-varying real-valued (just greater than 0 or 1) activations (outputs). Each connection can have a modifiable real-valued weight. Some of the nodes are called labeled nodes, some output nodes, and others hidden nodes. For supervised learning in a discrete-time setting, the training sequence of real-valued input vectors can be a sequence of activations of input nodes, one input vector at a time. At each time step, each non-input unit can calculate its current activation as a non-linear function of the weighted sum of the activations of all units from which it receives a connection. The system can explicitly activate some output units (independent of the incoming signal) at certain time steps.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be used to provide a self-organizing neural network, such as a Kohonen self-organizing neural network, for visualization of views of data, such as low-dimensional views of high-dimensional data. -Organizational neural networks can be used. A self-organizing neural network may apply competitive learning to a set of input data, for example, from an individual or from one or more sensors or other data inputs associated with the individual. In embodiments, self-organizing neural networks may be used to identify structure in data, such as unlabeled data, such as data from various unstructured sources, such as social media sources about individuals, where the source of the data is unknown (e.g., when data comes from various unknown or uncertain sources). Self-organizing neural networks can organize structures or patterns in data so that they can be recognized, analyzed, and labeled, such as identifying the structures as corresponding to individuals, disease conditions, health states, activity states, etc.

In embodiments, the methods and systems described in this application involving expert systems or self-organization capabilities may use recurrent neural networks, where connected units (e.g., neurons or nodes) form directed cycles and Likewise, bidirectional flow of data can be allowed. Such networks model or represent dynamic temporal behavior, such as that involved in dynamic systems such as the wide variety of disease conditions, health states, and biological systems described throughout this disclosure, such as bodies experiencing multiple different diseases or health states. It can be used to determine where dynamic system behavior involves complex interactions that an observer may wish to understand, diagnose, predict, control, treat and/or optimize. For example, recurrent neural networks can be used to predict an individual's state (e.g., maintenance state, health state, disease state, etc.), such as interacting with a system, performing an action, etc. In embodiments, recurrent neural networks may utilize inputs from other nodes and/or sensors and environments of various types described herein, such as social networks, home or work environments, healthcare environments, recreational or sports environments, etc. Internal memory can be used to process sequences of different data inputs from. In embodiments, recurrent neural networks may also be used for pattern recognition, such as to recognize people based on biomarkers, facial, voice or sound signatures, heat signatures, sets of feature vectors in an image, chemical signatures, etc. In a non-limiting example, a recurrent neural network may recognize changes or shifts in a human state by learning to classify shifts or changes from a training data set that consists of a stream of data from unstructured data sources, such as social media sources. You can.

In embodiments, the methods and systems described herein involving expert systems or self-organizing capabilities may include a set of independent neural networks (e.g., one of the various types described herein) mediated by an intermediary. A modular neural network can be used. Each independent neural network in a modular neural network can operate with separate inputs to accomplish subtasks that make up the task that the modular network as a whole is intended to perform. For example, a modular neural network can, once understood, identify some type of person, condition, or state by, e.g., one or more sensors providing input channels to the modular network and the RBF neural network to optimize the system, protocol, etc. It may include a recurrent neural network for pattern recognition to recognize whether something is being detected. The intermediary can accept inputs from each individual neural network, process them, and produce outputs for the modular neural network, such as appropriate control parameters, predictions of states, etc.

Combinations between any of the various neural network types described in this application, such as pairs, triplets, or larger combinations, are encompassed by this disclosure. This means that an expert system uses a single neural network to recognize patterns (e.g., patterns that indicate problem or fault conditions) and provides outputs that govern the system's autonomous control in response to the recognized conditions or patterns. ) may include a combination of using different neural networks to self-organize activities or workflows based on recognized patterns. This also means that expert systems use a neural network to classify an item (e.g., identify a machine, component, or mode of operation) and classify the item's state (e.g., defect state, operating state, expected state, maintenance, etc.). state, predicted state, etc.). Modular neural networks also allow expert systems to use one neural network to determine a state or situation (e.g., the state of a machine, process, workflow, storage system, network, data collector, etc.) and the processes that entail the state or situation (e.g. This may include situations where different neural networks are used to self-organize (for example, a data storage process, a network coding process, a network selection process, a data processing process, or other processes described herein).

In embodiments, the methods and systems described in this application involving expert systems or self-organization capabilities may use physical neural networks in which one or more hardware elements may be used to perform or simulate neural behavior. One or more hardware nodes may be configured to stream output data resulting from the activity of the neural network. Hardware nodes, which may include one or more chips, microprocessors, integrated circuits, programmable logic controllers, application-specific integrated circuits, field programmable gate arrays, etc., can be used to determine the speed of some portion of a neural network of any of the types described in this application; It may be provided to optimize input/output efficiency, energy efficiency, signal-to-noise ratio, or other parameters. Hardware nodes are hardware for acceleration of computations (e.g., dedicated processors to perform basic or more sophisticated calculations on input data to provide output, dedicated processors to filter or compress data, dedicated processors to decompress data). processor, a dedicated processor for compression of specific files or data types (e.g., for handling image data, video streams, acoustic signals, vibration data, thermal images, thermal maps, etc.). The physical neural network may be a data collector, an edge intelligence system, an adaptive intelligence system, a mobile data collector, an IoT monitoring system, or an expert system (optionally, may include a software-based neural network located on the data collector or remotely). may be implemented in other systems described in this application, including those that can be reconfigured by switching or routing inputs in various configurations, such as providing different neural network configurations within the system to handle different types of input (switching and configuring); You can. A physical or at least partially physical neural network may be used, e.g. to accelerate the input/output functions of one or more storage elements that feed data to or take data from the neural network, e.g. to store data within a machine, product, etc. , may include physical hardware nodes located in the storage system. A physical or at least partially physical neural network transmits data in, to, or from an environment, e.g., to accelerate input/output functions to one or more network nodes within a net, to accelerate relay functions, etc. To do so, it may include physical hardware nodes located on a network. In embodiments of physical neural networks, electrically tunable resistive materials can be used to emulate the function of neural synapses. In embodiments, physical hardware emulates neurons and software emulates neural networks between neurons. In embodiments, neural networks complement conventional algorithmic computers. They can be trained to perform appropriate functions without requiring arbitrary instructions, such as classification functions, optimization functions, pattern recognition functions, control functions, selection functions, evolution functions, etc.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use multilayer feedforward neural networks, e.g., for complex pattern classification of one or more items, phenomena, modes, states, etc. . In embodiments, multilayer feedforward neural networks may be trained by optimization techniques, such as genetic algorithms, to search a large, complex space of options to find an optimal or near-optimal global solution. For example, one or more genetic algorithms may be used to classify complex phenomena, e.g., modes involving complex interactions between entities (including interference effects, amplification effects, etc.), analysis of symptoms of entities, or diagnosis of conditions. Modes that involve non-linear phenomena, such as the impact of protocol interactions, which can make it difficult; modes that carry significant risks, such as when multiple simultaneous conditions occur, making root cause analysis difficult; and complex modes of individual behavior, such as Alternatively, it can be used to train a multilayer feedforward neural network to recognize states. In embodiments, a multi-layer feedforward neural network may be used to classify results from monitoring unstructured data, such as forming social media.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be used to address one or more remote sensing applications, e.g., in a variety of environments, including home and work environments, business environments, etc. To take input from sensors distributed throughout the human-inhabited environment, a feed-forward, backpropagation multilayer perceptron (MLP) neural network can be used. In an embodiment, an MLP neural network may be used for classification of the physical environment. This may include fuzzy classification.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use a structure-adaptive neural network, wherein the structure of the neural network adapts based on, e.g., rules, sensed conditions, situational parameters, etc. It can be. For example, if a neural network does not converge on a solution, such as classifying items or reaching a prediction, then when acting on a set of inputs after a certain amount of training, the neural network may By converting the data path from a unidirectional to a bidirectional data path, it can be modified from a feedforward neural network to a recurrent neural network. Structural adaptation refers to recognizing the occurrence of a critical value (e.g., lack of convergence to a solution within a given amount of time) or recognizing a phenomenon as requiring a different or additional structure (e.g., the ability for a system to dynamically or nonlinearly (e.g., to trigger adaptation upon the occurrence of a trigger, rule, or event), which may occur under the control of an expert system.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may include, for example, a multilayer perceptron (MLP), where there may be an input layer, an output layer, and one or more hidden layers connecting them. You can use autoencoder, autoassociator, or diabolo neural networks, which can be similar to neural networks. However, the output layer in an auto-encoder may have the same number of units as the input layer, where the goal of the MLP neural network may be to reconstruct its own input (rather than just emitting target values). Therefore, the autoencoder can operate as an unsupervised learning model. Autoencoders can be used for unsupervised learning of efficient coding, for example for dimensionality reduction, for learning generative models of data, etc. In embodiments, an auto-encoding neural network may be used to self-learn efficient network coding for transmission of data from or about an individual over one or more networks, which may include social networks. In embodiments, an auto-encoding neural network may be used to self-learn an efficient storage approach for storage of streams of analog sensor data from the environment.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may, in embodiments, include a probabilistic neural network (e.g., 4-layer), which may include a multi-layer (e.g., 4-layer) feedforward neural network. PNN) may be used, where the layers may include an input layer, a hidden layer, a pattern/summation layer, and an output layer. In an embodiment of the PNN algorithm, the parent probability distribution function (PDF) of each class may be approximated, for example, by a Pazen window and/or a non-parametric function. Then, using the PDF of each class, the class probability of the new input can be estimated, and Bayes' rule can be used, for example, to assign to the class with the highest posterior probability. PNN can implement a Bayesian network and can use statistical algorithms or analysis techniques such as kernel Fisher discriminant analysis technique. PNNs may be used for classification and pattern recognition in any of the wide range of embodiments disclosed in this application. In one non-limiting example, a probabilistic neural network may be used to predict defect conditions in a product or system based on a collection of data inputs from sensors and instruments on the engine.

In embodiments, the methods and systems described in this application involving expert systems or self-organization capabilities may use a time delay neural network (TDNN), which recognizes features independent of sequence position. May include a feedforward architecture for sequential data. In embodiments, delays are added to one or more inputs, or between one or more nodes, so that multiple data points (from separate points in time) are analyzed together to account for time shifts in the data. A time delay neural network can form part of a larger pattern recognition system, for example using a perceptron network. In embodiments, a TDNN may be trained with supervised learning, such as when connection weights are trained with backpropagation or under feedback. In embodiments, TDNNs may be used to process sensor data from separate streams, where time delays are used to temporally align data streams, e.g., to help understand patterns that accompany understanding of the various streams. It is used.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use convolutional neural networks (referred to in some cases as CNNs, ConvNets, shift invariant neural networks, or spatial invariant neural networks); , where the units are connected in a pattern similar to the visual cortex of the human brain. Neurons can respond to stimuli in a limited region of space called the receptive field. Receptive fields may partially overlap to collectively cover the entire (e.g., visual) field. The node response can be calculated mathematically, for example by a convolution operation, for example using a multilayer perceptron using minimal preprocessing. Convolutional neural networks can be used for recognition within image and video streams, such as recognizing individuals, recognizing markers of disease conditions, etc. This may include, for example, using camera systems deployed on mobile data collectors, such as drones or mobile robots, to recognize individuals in a crowd. In embodiments, convolutional neural networks may be used to provide recommendations based on data input, including sensor input and other contextual information. In embodiments, convolutional neural networks may be used to process input, such as natural language processing of instructions provided by one or more parties involved in workflow in the environment. In embodiments, a convolutional neural network may have multiple neurons (e.g., 100,000, 500,000 or more), many (e.g., 4, 5, 6 or more) layers, and many (e.g., millions). ) can be placed together with the parameter. A convolutional neural network can use one or more convolutional nets.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be used to, for example, recognize emergent phenomena (such as new types of conditions not previously understood in an individual or group of individuals). , regulatory feedback networks can be used.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use self-organizing maps (SOMs) involving unsupervised learning. A set of neurons can learn to map points in input space to coordinates in output space. The input space can have different dimensions and topology than the output space, and a SOM can preserve these while mapping phenomena into groups.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use learning vector quantization networks (LVQ). Prototypical representatives of a class can be parameterized in a distance-based classification technique, with an appropriate distance measure.

In embodiments, the methods and systems described in this application involving expert systems or self-organization capabilities may use echo state networks (ESNs), which may include recurrent neural networks with sparsely connected random hidden layers. The weights of the output neurons may be changed (e.g., the weights may be trained based on feedback). In embodiments, an ESN may be used to handle time series patterns, for example, to recognize patterns in the progression of a process.

In embodiments, methods and systems described herein involving expert systems or self-organizing capabilities may be used to predict or label each element of a sequence based on both past and future contexts of the element, e.g., using a finite number of values. A bidirectional, recurrent neural network (BRNN) can be used that uses sequences (e.g., voltage values from sensors). This can be done by adding the outputs of two RNNs, such that one processes the sequence from left to right and the other processes the sequence from right to left. The combined output is a prediction of the target signal, such as that provided by a teacher or supervisor. Bidirectional RNNs can be combined with short- and long-term memory RNNs.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use hierarchical RNNs that connect elements in various ways to decompose hierarchical behavior into useful subprograms, for example. . In embodiments, a hierarchical RNN may be used to manage one or more hierarchical templates for data collection in social networks, value chain environments, etc.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use probabilistic neural networks, which may introduce random fluctuations into the network. These random fluctuations can be considered a form of statistical sampling, such as Monte Carlo sampling or other statistical sampling techniques.

In embodiments, the methods and systems described herein involving expert systems or self-organizing capabilities may use genetic scale recurrent neural networks. In these embodiments, a RNN (often a LSTM) may be used, where the series can be decomposed into a number of scales where every scale gives the primary length between two consecutive points. The first scale consists of a normal RNN, the second scale consists of all points separated by two indices, and so on. Nth-order RNN connects the first node and the last node. The output from all the various scales can be treated as a committee of members, and the associated scores can be used genetically for the next iteration.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use a Committee of Machines (CoM) comprising a collection of different neural networks that “vote” together on a given example. Because neural networks may suffer from local minima starting with the same architecture and training, using randomly different initial weights often gives different results. CoM tends to stabilize results.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be associated with association, such as those involving extensions of machine committees combining multiple feedforward neural networks and k-nearest neighbor techniques. A neural network (ASNN) can be used. This allows using the correlation between ensemble responses as a measure of distance among the analyzed cases for kNN. This corrects bias in the neural network ensemble. An associative neural network can have a memory that can match the training set. When new data becomes available, the network immediately improves its prediction ability and provides data approximation without retraining (self-learning). Another important feature of ASNN may be the possibility to interpret neural network results by analysis of correlations between data instances in the space of the model.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use an Instantaneous Training Neural Network (ITNN), where the weights of the hidden and output layers are mapped directly from training vector data.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use spiking neural networks that can explicitly take into account the timing of inputs. Network inputs and outputs can be represented as a series of spikes (such as delta functions or more complex shapes). SNNs can process information in the time domain (e.g., signals that change over time, such as signals involving an individual's dynamic behavior, disease conditions, health status, etc.). These can be implemented as circular networks.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use dynamic neural networks that address nonlinear multivariate behavior and include learning of time-dependent behavior such as transients and delay effects. Transient phenomena may include progressive state behavior.

In embodiments, cascade correlation may be used as an architecture and supervised learning algorithm to supplement the adjustment of weights in networks of fixed topology. Cascade-correlation can start with a minimal network, and then automatically train and add new hidden units one by one, creating a multi-layer structure. Once a new hidden unit is added to the network, its input weights can be frozen. This unit then becomes a permanent feature detector in the network, available to generate output or generate other more complex feature detectors. Cascade-correlated architectures can learn quickly, determine their own size and topology, maintain the structure they build even when the training set changes, and do not require backpropagation.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use, for example, neuro-fuzzy networks involving a fuzzy interference system in the body of an artificial neural network. Depending on the type, multiple layers can simulate the processes involved in fuzzy inference, such as fuzzification, inference, aggregation, and defuzzification. We embed the fuzzy system in the general structure of a neural network, exploiting the use of available training methods to find the parameters of the fuzzy system.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be used in a configurational pattern generation network (CPPN), such as a variant of an association neural network (ANN), that differentiates a set of activation functions and how they are applied. ) can be used. Typical ANNs often contain only sigmoid functions (and sometimes Gaussian functions). A PPN can include both types of functionality. Additionally, CPPN can be applied over the entire space of possible inputs, so they can represent a complete image. Because they are compositions of functions, CPPN encodes images at virtually infinite resolution and can be sampled for a particular display at whatever resolution may be optimal. This type of network can add new patterns without retraining. In embodiments, the methods and systems described herein involving expert systems or self-organization capabilities create a specific memory structure that assigns each new pattern to an orthogonal plane, e.g., using adjacently connected hierarchical arrays. By doing so, a one-shot associative memory network can be used.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use, for example, hierarchical temporal memory (HTM) neural networks that incorporate the structural and algorithmic properties of the neocortex. HTM can use biomimetic models, such as those based on memory prediction. HTM can be used to discover and infer high-level causes of observed input patterns and sequences.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use holographic associative memory (HAM) neural networks, which may include analog, correlation-based, associative, stimulus-response systems. there is. Information can be mapped to complex phase orientations. Memory can be effective in associative memory tasks, generalization, and pattern recognition with mutable attention.

The neural network described above may have various nodes or neurons, which may perform various functions on input, such as input received from sensors or other data sources, including other nodes. A function may involve weights, features, feature vectors, etc. Neurons may include perceptrons, neurons that mimic biological functions (such as the human senses of touch, sight, taste, hearing, and smell). Continuous neurons, such as sigmoidal activation, can be used in the context of various types of neural networks, such as those involving backpropagation.

In many embodiments, an expert system or neural network may be trained, such as by a human operator or supervisor, or based on a data set, model, etc. Training may include one or more training data sets representing values such as sensor data, event data, parameter data, and other types of data (including many of the types described throughout this disclosure), as well as the results of the process, It may include presenting one or more indicators of a result, such as the result of a calculation, the result of an event, the result of an activity, etc., to the neural network. Training a neural network to optimize one or more systems based on one or more optimization approaches such as Bayesian approach, parametric Bayes classifier approach, k-nearest-neighbor classifier approach, iterative approach, interpolation approach, Pareto optimization approach, algorithmic approach, etc. This may include training in optimization, such as training . Feedback may be provided in a process of variation and selection, such as using a genetic algorithm to evolve one or more solutions based on feedback over a series of rounds.

In embodiments, receiving data streams and other inputs collected from one or more industrial environments (e.g., by a mobile data collector) and transmitted to a cloud platform over one or more networks, including using network coding to provide efficient transmission. Multiple neural networks can be deployed on a cloud platform. In the cloud platform, a plurality of different neural networks of several types (including modular form, structural adaptive form, hybrid, etc.) are performed to perform prediction, classification, and control functions, optionally using massively parallel computing power, and the present disclosure It may be used to provide other outputs as described with respect to the expert system disclosed throughout. Different neural networks can be selected, for example by an expert system, for a particular task involved in a given situation, workflow, environmental process, system, etc., so that an appropriate type of neural network with appropriate input set, weights, node types and functions, etc. They can be structured to compete with each other (optionally including the use of evolutionary algorithms, genetic algorithms, etc.).

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may provide information in one direction, e.g., from a data input such as an analog sensor located on or proximate to an industrial machine; You can use a feedforward neural network that moves the output through a series of neurons, or nodes. Data can move from input nodes to output nodes, and optionally through one or more hidden nodes without looping. In embodiments, a feedforward neural network may be composed of various types of units, such as binary McCulloch-Pitts neurons, the simplest of which is a perceptron.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be used in some situations involving interpolation in a multidimensional space (e.g., when interpolation helps optimize a multidimensional function). RBF (which may be desirable, e.g., when optimizing data markets as described in this application, when optimizing the efficiency or output of power generation systems, factory systems, etc., or in other situations involving multiple dimensions) A radial basis function (radial basis function) neural network can be used. In an embodiment, each neuron within the RBF neural network stores examples from the training set as “prototypes.” The linearity involved in the functioning of this neural network gives RBF the advantage of not typically suffering from local minima or maxima problems.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may utilize radial basis function (RBF) neural networks, such as those using a distance to centroid criterion (e.g., a Gaussian function). there is. Radial basis functions can be applied as a replacement for hidden layers (such as sigmoidal hidden layer propagation) in multilayer perceptrons. An RBF network may have two layers, such that inputs are mapped onto each RBF in a hidden layer. In an embodiment, the output layer may include, for example, a linear combination of hidden layer values representing the average predicted output. The output layer values can provide the same or similar output to that of a regression model in statistics. In a classification problem, the output layer may be a sigmoid function of a linear combination of hidden layer values representing the posterior probability. In both cases, performance is often improved by shrinkage techniques, such as ridge regression in classical statistics. This corresponds to the a priori confidence in small parameter values (and therefore smooth output functions) in the Bayesian framework. RBF networks can avoid local minima because the only parameter adjusted in the learning process is the linear mapping from the hidden layer to the output layer. Linearity ensures that the error surface is quadratic and therefore has a single minimum. In regression problems, this can be found in one matrix operation. In classification problems, the fixed nonlinearity introduced by the sigmoid output function can be handled using an iteratively reweighted least squares function, etc.

RBF networks can use kernel methods such as support vector machines (SVM) and Gaussian processes (where RBF is the kernel function). A non-linear kernel function can be used to project the input data into a space where the learning problem can be solved using a linear model.

In an embodiment, an RBF neural network may include an input layer, a hidden layer, and a summation layer. In the input layer, one neuron appears in the input layer for each predictor. For categorical variables, N-1 neurons are used, where N is the number of categories. The input neuron may, in embodiments, normalize the range of values by subtracting the median and dividing by the interquartile range. The input neuron can then feed the value to each neuron in the hidden layer. In the hidden layer, a variable number of neurons can be used (determined by the training process). Each neuron can be constructed with a radial basis function centered on a point that has as many dimensions as the number of predictors. The spread (eg, radius) of the RBF function may be different for each dimension. Center and spread can be determined by training. When presented with a vector of input values from the input layer, the hidden neuron can calculate the Euclidean distance of the test case from the neuron's center point and then apply the RBF kernel function to this distance, for example using the diffusion value. there is. The resulting value can then be passed to the summation layer. In the summation layer, the value originating from a neuron in the hidden layer can be multiplied by the weight associated with the neuron and added to the weighted values of other neurons. This sum becomes the output. For classification problems, one output is generated (with a separate set of weights and summing units) for each target category. The value output for a category is the probability that the case being evaluated has that category. In the training of RBF, the number of neurons in the hidden layer, the coordinates of the centroid of each hidden layer function, the spread of each function in each dimension, and the weights applied to the output when it is passed to the summation layer. Various parameters can be determined. Training can be used by clustering algorithms (such as k-means clustering), evolutionary approaches, etc.

In embodiments, a recurrent neural network may have time-varying real-valued (just greater than 0 or 1) activations (outputs). Each connection can have a modifiable real-valued weight. Some of the nodes are called labeled nodes, some output nodes, and others hidden nodes. For supervised learning in a discrete-time setting, the training sequence of real-valued input vectors can be a sequence of activations of input nodes, one input vector at a time. At each time step, each non-input unit can calculate its current activation as a non-linear function of the weighted sum of the activations of all units from which it receives a connection. The system can explicitly activate some output units (independent of the incoming signal) at certain time steps.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be used to provide a self-organizing neural network, such as a Kohonen self-organizing neural network, for visualization of views of data, such as low-dimensional views of high-dimensional data. -Organizational neural networks can be used. A self-organizing neural network can apply competitive learning to a set of input data, such as from an industrial machine or from one or more sensors or other data inputs associated with the industrial machine. In embodiments, a self-organizing neural network may be used to detect, for example, vibration, acoustic, or other analog sensors in an industrial environment where the source of the data is unknown (e.g., the vibration may originate from any of a range of unknown sources). In data sensed from a range of, it can be used to identify structures in the data, such as unlabeled data. Self-organizing neural networks allow them to be recognized, analyzed, and labeled, such as identifying structures as corresponding to vibrations induced by the movement of a floor, or acoustic signals generated by the high-frequency rotation of a more or less distant machine's shaft. You can organize structures or patterns in your data so that you can

In embodiments, the methods and systems described in this application involving expert systems or self-organization capabilities may use recurrent neural networks, e.g., where connected units (e.g., neurons or nodes) form directed cycles. In this case, bidirectional flow of data can be allowed. These networks are involved in dynamic systems, including the wide variety of industrial machines and devices described throughout this disclosure, such as power generation machines, robotic manufacturing systems, retrofit systems, etc., operating at variable speeds or frequencies under variable conditions with variable inputs. It can be used to model or represent dynamic temporal behavior, such as where dynamic system behavior involves complex interactions that operators may wish to understand, predict, control, and/or optimize. For example, recurrent neural networks can be used, for example, to predict the state (e.g., maintenance state, fault state, operating state, etc.) of an industrial machine performing a dynamic process or action. In embodiments, a recurrent neural network may use internal memory to process sequences of inputs from other nodes and/or sensors and other data inputs, such as those described herein, of the various types described herein. In embodiments, recurrent neural networks may also be used for pattern recognition, such as to recognize industrial machines based on sound signatures, heat signatures, sets of feature vectors in an image, chemical signatures, etc. In a non-limiting example, a recurrent neural network can be used to classify shifts in turbines, generators, and motors by learning to classify shifts from a training data set consisting of a stream of data from three-axis vibration sensors and/or acoustic sensors applied to one or more of these machines. , a shift (eg, gear shift) in the operation mode of a compressor, etc. can be recognized.

In embodiments, the methods and systems described herein involving expert systems or self-organizing capabilities may include a set of independent neural networks (e.g., one of the various types described herein) mediated by an intermediary. A modular neural network can be used. Each independent neural network in a modular neural network can operate with separate inputs to accomplish subtasks that make up the task that the modular network as a whole is intended to perform. For example, a modular neural network may be a recurrent neural network for pattern recognition, such as for recognizing what type of industrial machine is being sensed by one or more sensors that are provided as input channels to the modular network, and then , may include an RBF neural network to optimize the behavior of the machine. The intermediary can accept inputs from each individual neural network, process them, and produce outputs for the modular neural network, such as appropriate control parameters, predictions of states, etc.

Combinations between any of the various neural network types described in this application, such as pairs, triplets, or larger combinations, are encompassed by this disclosure. This means that an expert system uses a single neural network to recognize patterns (for example, patterns that indicate problem or fault conditions) and provides outputs that govern the system's autonomous control in response to the recognized conditions or patterns. ) may include a combination of using different neural networks to self-organize activities or workflows based on recognized patterns. This also allows the expert system to use one neural network to classify the item (e.g., identify the machine, component, or mode of operation) and the state of the item (e.g., defect state, operating state, expected state, maintenance state, etc.). ) may include a combination of using different neural networks to predict. Modular neural networks can also be defined when an expert system uses a single neural network to determine a state or situation (e.g., state of a machine, process, workflow, market, storage system, network, data collector, etc.), and entails the state or situation. A process (e.g., a data storage process, a network coding process, a network selection process, a data market process, a power generation process, a manufacturing process, an improvement process, a digging process, a boring process, or any of the methods described in this application) may include situations where different neural networks are used to self-organize different processes).

In embodiments, the methods and systems described in this application involving expert systems or self-organization capabilities may use physical neural networks in which one or more hardware elements are used to perform or simulate neural behavior. In embodiments, one or more hardware neurons may stream voltage values representing analog vibration sensor data voltage values, compute velocity information from analog sensor inputs representing acoustic, vibration, or other data, and calculate velocity information from analog sensor inputs representing acoustic, vibration, or other data. It may be configured to calculate acceleration information from input, etc. One or more hardware nodes may be configured to stream output data resulting from the activity of the neural network. Hardware nodes, which may include one or more chips, microprocessors, integrated circuits, programmable logic controllers, application-specific integrated circuits, field programmable gate arrays, etc., can be used to determine the speed of some portion of a neural network of any of the types described in this application; It may be provided to optimize input/output efficiency, energy efficiency, signal-to-noise ratio, or other parameters. Hardware nodes are hardware for acceleration of computations (e.g., dedicated processors to perform basic or more sophisticated calculations on input data to provide output, dedicated processors to filter or compress data, dedicated processors to decompress data). processor, a dedicated processor for compression of specific files or data types (e.g., for handling image data, video streams, acoustic signals, vibration data, thermal images, thermal maps, etc.). A physical neural network may be configured to have different neural network configurations within the data collector to handle different types of input (optionally, with switching and configuration under the control of an expert system, which may include a software-based neural network located on the data collector or remotely). It may be implemented in a data collector, such as a mobile data collector described herein, including one that may be reconfigured by switching or routing inputs in various configurations, as provided herein. A physical or at least partially physical neural network stores data, e.g. within an industrial machine or in an industrial environment, for example to accelerate input/output functions for one or more storage elements that feed data to or take data from the neural network. To do so, it may include a physical hardware node located in the storage system. A physical or at least partially physical neural network transmits data, e.g., within an environment, to or from an industrial environment, e.g., to accelerate input/output functions to one or more network nodes within a net, to accelerate relay functions, etc. To do so, it may include physical hardware nodes located on a network. In embodiments of physical neural networks, electrically tunable resistive materials can be used to emulate the function of neural synapses. In embodiments, physical hardware emulates neurons and software emulates neural networks between neurons. In embodiments, neural networks complement conventional algorithmic computers. They are versatile and can be trained to perform appropriate functions without requiring arbitrary instructions, such as classification functions, optimization functions, pattern recognition functions, control functions, selection functions, evolution functions, etc.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use multi-layer feedforward neural networks, e.g., for complex pattern classification of one or more items, phenomena, modes, states, etc. there is. In embodiments, multilayer feedforward neural networks may be trained by optimization techniques, such as genetic algorithms, to search a large, complex space of options to find an optimal or near-optimal global solution. For example, one or more genetic algorithms may be used to determine, for example, the effects of variable speed shafts, which can make analysis of modes, vibrations and other signals that involve complex interactions between machines (including interference effects, resonance effects, etc.) difficult to analyze. To recognize the complex operating modes of industrial machines, to classify complex phenomena, such as modes involving non-linear phenomena such as modes involving significant faults such as multiple simultaneous faults occurring, making root cause analysis difficult, etc. It can be used to train multilayer feedforward neural networks. In an embodiment, a multi-layer feedforward neural network can be used in industrial applications, such as monitoring an internal set of components within a housing such as motor components, pumps, valves, fluid handling components, etc., in refrigeration systems, retrofit systems, reactor systems, catalyst systems, etc. It can be used to classify results from ultrasonic or acoustic monitoring of a machine.

In embodiments, the methods and systems described herein involving expert systems or self-organizing capabilities address one or more remote sensing applications, e.g., to take input from sensors distributed throughout a variety of industrial environments. To do this, a feedforward, backpropagation multilayer perceptron (MLP) neural network can be used. In embodiments, the MLP neural network may be used to solve problems such as mining environments, exploration environments, including classification of geological structures (including subsurface features and above-ground features), classification of materials (including fluids, minerals, metals, etc.), and other problems. It can be used for classification of physical environments such as drilling environments. This may include fuzzy classification.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use a structure-adaptive neural network, wherein the structure of the neural network adapts based on, e.g., rules, sensed conditions, situational parameters, etc. do. For example, if a neural network does not converge on a solution, such as classifying items or reaching a prediction, then when acting on a set of inputs after a certain amount of training, the neural network may By converting the data path from a unidirectional to a bidirectional data path, it can be modified from a feedforward neural network to a recurrent neural network. Structural adaptation refers to recognizing the occurrence of a critical value (e.g., lack of convergence to a solution within a given amount of time) or recognizing a phenomenon as requiring a different or additional structure (e.g., the ability for a system to dynamically or nonlinearly (recognizing that something is changing in a certain way) can occur under the control of an expert system, for example, to trigger adaptation upon the occurrence of a trigger, rule, or event. In one non-limiting example, when the expert system receives an indication that a continuously variable transmission is being used to drive generators, turbines, etc. in the system being analyzed, the expert system can transform from a simple neural network structure such as a feedforward neural network to a recurrent neural network, a convolutional neural network, etc. You can switch to more complex neural network structures, such as neural networks.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may include, for example, a multilayer perceptron (“MLP”), where there may be an input layer, an output layer, and one or more hidden layers connecting them. ") You can use autoencoder, autoassociator, or diabolo neural networks, which can be similar to neural networks. However, the output layer in an auto-encoder can have the same number of units as the input layer, where the goal of the MLP neural network is to reconstruct its own input (rather than just emitting target values). Therefore, the autoencoder can operate as an unsupervised learning model. Autoencoders can be used for unsupervised learning of efficient coding, for example for dimensionality reduction, for learning generative models of data, etc. In embodiments, an auto-encoding neural network may be used to self-learn efficient network coding for transmission of analog sensor data from industrial machines over one or more networks. In an embodiment, an auto-encoding neural network can be used to self-learn an efficient storage approach for storage of streams of analog sensor data from an industrial environment.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may, in embodiments, include a probabilistic neural network (e.g., 4-layer), which may include a multi-layer (e.g., 4-layer) feedforward neural network. “PNN”) may be used, where the layers may include an input layer, a hidden layer, a pattern/summation layer, and an output layer. In an embodiment of the PNN algorithm, the parent probability distribution function (PDF) of each class may be approximated, for example, by a Pazen window and/or a non-parametric function. Then, using the PDF of each class, the class probability of the new input is estimated, and Bayes' rule can be used, for example, to assign it to the class with the highest posterior probability. PNN can implement a Bayesian network and can use statistical algorithms or analysis techniques such as kernel Fisher discriminant analysis technique. PNNs may be used for classification and pattern recognition in any of the wide range of embodiments disclosed in this application. In one non-limiting example, a probabilistic neural network may be used to predict fault conditions for an engine based on a collection of data inputs from sensors and instruments on the engine.

In embodiments, the methods and systems described in this application involving expert systems or self-organization capabilities may use a time delay neural network (TDNN), which recognizes features independent of sequence position. May include a feedforward architecture for sequential data. In embodiments, delays are added to one or more inputs, or between one or more nodes, so that multiple data points (from separate points in time) are analyzed together to account for time shifts in the data. A time delay neural network can form part of a larger pattern recognition system, for example using a perceptron network. In embodiments, a TDNN may be trained with supervised learning, such as when connection weights are trained with backpropagation or under feedback. In embodiments, a TDNN may be used to process sensor data from distinct streams, such as a stream of velocity data, a stream of acceleration data, a stream of temperature data, a stream of pressure data, etc., where the time delay is present in the various streams, e.g. It is used to align data streams temporally to help understand patterns that entail understanding (for example, increases in pressure and acceleration as an industrial machine overheats).

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use convolutional neural networks (referred to in some cases as CNNs, ConvNets, shift invariant neural networks, or spatial invariant neural networks); , where the units are connected in a pattern similar to the visual cortex of the human brain. Neurons can respond to stimuli in a limited region of space called the receptive field. Receptive fields may partially overlap to collectively cover the entire (e.g., visual) field. The node response can be calculated mathematically, for example by a convolution operation using a multilayer perceptron using minimal preprocessing. Convolutional neural networks can be used to recognize types of machines in large-scale environments, for example, using camera systems deployed on mobile data collectors such as drones or mobile robots for recognition within image and video streams. In embodiments, convolutional neural networks may be used to provide recommendations based on data input, including sensor input and other contextual information, such as recommending a route for a mobile data collector. In embodiments, convolutional neural networks may be used to process input, such as natural language processing of instructions provided by one or more parties involved in workflow in the environment. In embodiments, a convolutional neural network may have multiple neurons (e.g., 100,000, 500,000 or more), many (e.g., 4, 5, 6 or more) layers, and many (e.g., millions). ) can be placed together with the parameter. A convolutional neural network can use one or more convolutional nets.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be used to regulate, e.g., to recognize emergent phenomena (such as new types of defects not previously understood in industrial environments). You can use a feedback network.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use self-organizing maps (“SOMs”) involving unsupervised learning. A set of neurons can learn to map points in input space to coordinates in output space. The input space can have different dimensions and topology than the output space, and a SOM can preserve these while mapping phenomena into groups.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use learning vector quantization neural networks (“LVQ”). Prototypical representatives of a class can be parameterized in a distance-based classification technique, with an appropriate distance measure.

In embodiments, the methods and systems described in this application involving expert systems or self-organization capabilities may use echo state networks (“ESNs”), which may include recurrent neural networks with sparsely connected random hidden layers. . The weights of the output neurons may be changed (e.g., the weights may be trained based on feedback). In embodiments, ESN may be used to handle time series patterns, such as recognizing patterns of events associated with gear shifts in industrial turbines, generators, etc., for example.

In embodiments, methods and systems described in this application involving expert systems or self-organizing capabilities may be used to predict or label each element of a sequence based on both past and future contexts of the element, e.g., using a finite number of values. A bidirectional, recurrent neural network (“BRNN”) that uses a sequence (e.g., voltage values from a sensor) can be used. This can be done by adding the outputs of two RNNs, such that one processes the sequence from left to right and the other processes the sequence from right to left. The combined output is a prediction of the target signal, such as that provided by a teacher or supervisor. Bidirectional RNNs can be combined with short- and long-term memory RNNs.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use hierarchical RNNs that connect elements in various ways to decompose hierarchical behavior into useful subprograms, for example. . In embodiments, a hierarchical RNN may be used to manage one or more hierarchical templates for data collection in an industrial environment.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use probabilistic neural networks, which may introduce random fluctuations into the network. These random fluctuations can be considered as a form of statistical sampling, such as Monte Carlo sampling.

In embodiments, the methods and systems described herein involving expert systems or self-organizing capabilities may use genetic scale recurrent neural networks. In these embodiments, RNNs (often LSTMs) are used in which the series is decomposed into a number of scales, where every scale gives the linear length between two consecutive points. The first scale consists of a normal RNN, the second scale consists of all points separated by two indices, and so on. Nth-order RNN connects the first node and the last node. The output from all the various scales can be treated as a committee of members, and the associated scores can be used genetically for the next iteration.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use a Committee of Machines ("CoM") comprising a collection of different neural networks that "vote" together on a given example. . Because neural networks may suffer from local minima starting with the same architecture and training, using randomly different initial weights often gives different results. CoM tends to stabilize results.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may involve, for example, extensions of machine committees combining multiple feedforward neural networks and k-nearest neighbor techniques. A neural network (“ASNN”) can be used. This allows using the correlation between ensemble responses as a measure of distance among the analyzed cases for kNN. This corrects for bias in the neural network ensemble. An associative neural network can have a memory that can match the training set. When new data becomes available, the network immediately improves its prediction ability and provides data approximation without retraining (self-learning). Another important feature of ASNN is the possibility to interpret neural network results by analysis of correlations between data instances in the space of the model.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use an Instantaneous Training Neural Network (“ITNN”), where the weights of the hidden and output layers are mapped directly from training vector data. do.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use spiking neural networks that can explicitly take into account the timing of inputs. Network inputs and outputs can be represented as a series of spikes (such as delta functions or more complex shapes). SNNs can process information in the time domain (for example, signals that change over time, such as signals accompanying the dynamic behavior of industrial machinery). These are often implemented as circular networks.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use dynamic neural networks that address nonlinear multivariate behavior and include learning of time-dependent behavior such as transients and delay effects. Transients may include the behavior of shifting industrial components, such as the variable speed of a rotating shaft or other rotating component.

In embodiments, cascade correlation may be used as an architecture and supervised learning algorithm to supplement the adjustment of weights in networks of fixed topology. Cascade-correlation can start with a minimal network, and then automatically train and add new hidden units one by one, creating a multi-layer structure. Once a new hidden unit is added to the network, its input weights can be frozen. This unit then becomes a permanent feature detector in the network, available to generate output or generate other more complex feature detectors. Cascade-correlated architectures can learn quickly, determine their own size and topology, maintain the structure they build even when the training set changes, and do not require backpropagation.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use, for example, neuro-fuzzy networks involving a fuzzy inference system on the body of an artificial neural network. Depending on the type, multiple layers can simulate the processes involved in fuzzy inference, such as fuzzification, inference, aggregation, and defuzzification. We embed the fuzzy system in the general structure of a neural network, exploiting the use of available training methods to find the parameters of the fuzzy system.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may be comprised of a set of activation functions and how they are applied and how they are applied and configured pattern generation networks, such as variants of Associative Neural Networks (“ANNs”). "CPPN") can be used. Typical ANNs often contain only sigmoid functions (and sometimes Gaussian functions), but CPPNs can contain both types of functions, and so on. Additionally, CPPN can be applied over the entire space of possible inputs, so they can represent a complete image. Because they are compositions of functions, CPPN encodes images at virtually infinite resolution and can be sampled for a particular display at the optimal resolution.

This type of network can add new patterns without retraining. In embodiments, the methods and systems described herein involving expert systems or self-organizing capabilities create a specific memory structure that assigns each new pattern to an orthogonal plane, e.g., using adjacently connected hierarchical arrays. By doing so, a one-shot associative memory network can be used.

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use hierarchical temporal memory (“HTM”) neural networks, e.g., involving the structural and algorithmic properties of the neocortex. there is. HTM can use a biomimetic model based on memory prediction theory. HTM can be used to discover and infer high-level causes of observed input patterns and sequences.

In embodiments, the methods and systems described in this application involving expert systems or self-organization capabilities include holographic associative memory (“HAM”) neural networks, which may include analog, correlation-based, associative, stimulus-response systems. You can use it. Information can be mapped to complex phase orientations. Memory is effective for associative memory tasks, generalization, and pattern recognition with mutable attention.

INTELLIGENCE SYSTEM

104 illustrates an example intelligent services system 8800 (also referred to as “intelligent services”) in accordance with some embodiments of the present disclosure. In an embodiment, intelligent services 8800 provides a framework for providing intelligent services to one or more intelligent service clients 8836. In some embodiments, the intelligent services 8800 framework can be adapted to be at least partially replicated in each intelligent client 8836 (e.g., a VCN control tower and/or various VCN entities). In such embodiments, individual clients 8836 may include some or all of the capabilities of intelligent service 8800, whereby intelligent service 8800 adapts to specific functions performed by the intelligent client's subsystems. do. Additionally or alternatively, in some embodiments, intelligence services 8800 are implemented as a set of microservices such that different intelligence clients 8836 can utilize intelligence services 8800 through one or more APIs exposed to the intelligence clients. It can be. In this embodiment, intelligent services 8800 can be configured to perform various types of intelligent services that can be adapted for different intelligent clients 8836. In either of these configurations, an intelligence service client 8836 may provide an intelligence request to the intelligence service 8800, where the request is for a specific intelligence task (e.g., decision, recommendation, report, command, classification, predictions, training actions, NLP requests, etc.). In response, intelligent service 8800 executes the requested intelligent task and returns a response to intelligent service client 8836. Additionally or alternatively, in some embodiments, intelligence services 8800 may use one or more specialized chips configured to provide AI-assisted microservices such as image processing, diagnostics, location and orientation, chemical analysis, data processing, etc. It can be implemented. Examples of AI-enabled chips are discussed elsewhere in this disclosure.

In an embodiment, intelligent service 8800 may include an intelligent service controller 8802 and an artificial intelligence (AI) module 8804. In an embodiment, artificial intelligence service 8800 receives intelligence requests from intelligent service clients 8836 and any necessary data to process requests from intelligent service clients 8836. In response to the request and specific data, one or more associated artificial intelligence modules 8804 perform intelligent tasks and output “intelligence responses.” Examples of intelligence module 8804 responses include decisions (e.g., control commands, proposed actions, machine-generated text, etc.), predictions (e.g., predicted meaning of text snippets, predicted outcomes associated with proposed actions, etc.) , predicted fault conditions, etc.), classification (e.g., classification of objects in images, classification of vocal sounds, classified fault conditions based on sensor data, etc.), and/or other suitable outputs of artificial intelligence systems. You can.

In an embodiment, artificial intelligence module 8804 includes ML module 8812, rule-based module 8828, analytics module 8818, RPA module 8816, digital twin module 8820, machine vision module 8822, It may include an NLP module 8824, and/or a neural network module 8814. It is understood that the foregoing are non-limiting examples of artificial intelligence modules, and that some of the modules may be included or utilized by other artificial intelligence modules. For example, NLP module 8824 and machine vision module 8822 may utilize different neural networks that are part of neural network module 8814 in performing their respective functions.

It should be further noted that in some scenarios, the artificial intelligence module 8804 itself may also be an intelligence client 8836. For example, rule-based intelligence module 8828 may request intelligence tasks from ML module 8812 or neural network (F41) module 8814, such as requesting classification of objects and/or motion of objects that appear in a video. there is. In this example, rule-based intelligence module 8828 may be an intelligence service client 8836 that uses the classification to decide whether to take a specified action. In another example, machine vision module 8822 may request a digital twin of a specified environment from digital twin module 8820, so that ML module 8812 can use the digital twin as features to train a machine learning model that is trained for the specific environment. You can request specific data from the twin.

In embodiments, an intelligence task may require certain types of data to respond to the request. For example, a machine vision task is to classify objects that appear in one or more images (and potentially other data) in an image or set of images (position of the item, presence of a face, symbols or commands, expressions, parameters of motion, It requires determining features within a set of images (such as changes in state, etc.). In other examples, NLP tasks require audio of dialogue and/or text data (and potentially other data) to determine meaning or other elements of the dialogue and/or text. In another example, an AI-based control task (e.g., decisions about a robot's movement) may require environmental data (e.g., maps, coordinates, images of known obstacles) to make decisions about how to control the robot's motion. etc.) and/or motion planning may be required. In a platform-level example, analytics-based reporting tasks may require data from multiple different databases to generate reports. Accordingly, in embodiments, tasks that may be performed by intelligent service 8800 may require or benefit from specific intelligent service input 8832. In some embodiments, intelligent service 8800 may be configured to receive and/or request specific data from intelligent service input 8832 to perform each intelligent task. Additionally or alternatively, the requesting intelligence service client 8836 may provide specific data in the request. For example, intelligence service 8800 may expose one or more APIs to intelligence clients 8836, whereby requesting clients 8836 provide specific data in a request via the API. Examples of intelligent service inputs may include, but are not limited to, sensors providing sensor data, video streams, audio streams, databases, data feeds, human input, and/or other suitable data.

In an embodiment, intelligence module 8804 includes and provides access to an ML module 8812 that may be integrated into or accessed by one or more intelligence clients 8836. In an embodiment, the ML module 8812 provides services to the intelligence services client 8836, such as training ML models, utilizing ML models, enriching ML models, performing various clustering techniques, feature extraction, etc. Provide machine-based learning capabilities, features, functions, and algorithms for use by In an example, machine learning module 8812 may provide machine learning computing, data storage, and feedback infrastructure to the simulation system (e.g., as described above). The machine learning module 8812 may also operate in cooperation with other modules, such as the rule-based module 8828, the machine vision module 8822, the RPA module 8816, etc.

Machine learning module 8812 is one for performing analysis, simulation, decision-making, and predictive analytics related to data processing, data analysis, simulation generation, and simulation analysis of one or more components or subsystems of intelligent service client 8836. The above machine learning model can be defined. In embodiments, a machine learning model is an algorithm and/or statistical model that performs a specific task without using explicit instructions and instead relying on patterns and inferences. Machine learning models build one or more mathematical models based on training data to make predictions and/or decisions without being explicitly programmed to perform a specific task. In example implementations, machine learning models may perform classification, prediction, regression, clustering, anomaly detection, recommendation generation, and/or other tasks.

In embodiments, machine learning models may perform various types of classification based on input data. Classification is a predictive modeling problem in which class labels are predicted for given examples of input data. For example, machine learning models can perform binary classification, multi-class, or multi-label classification. In an embodiment, a machine learning model may output “confidence scores” that indicate the respective confidence levels associated with the classification of the input into each class. In embodiments, confidence scores may be compared to one or more thresholds to render discrete category predictions. In an embodiment, only a certain number of classes (e.g., one) with the relatively largest confidence score may be selected to render a discrete category prediction.

In embodiments, a machine learning model may output a probabilistic classification. For example, a machine learning model can predict a probability distribution over a set of classes, given sample input. Therefore, rather than outputting only the most likely class to which the sample input should belong, the machine learning model can output, for each class, the probability that the sample input belongs to that class. In an embodiment, the probability distribution for all possible classes may sum to 1. In embodiments, a softmax function, or other type of function or hierarchy, may be used to convert a set of real values each associated with a possible class into a set of real values in the range (0, 1) that sums to 1. In embodiments, the probabilities provided by the probability distribution may be compared to one or more thresholds to render discrete category predictions. In embodiments, only a certain number of classes (e.g., one) with the relatively greatest prediction probability may be selected to render discrete category predictions.

In embodiments, a machine learning model may perform regression to provide output data in the form of continuous numeric values. By way of example, a machine learning model may perform linear regression, polynomial regression, or non-linear regression. As described, in embodiments, the softmax function or other function or layer squashes the set of real values each associated with two or more possible classes into a set of real values in the range (0, 1) that sums to 1. can be used to For example, machine learning models can perform linear regression, polynomial regression, or nonlinear regression. As an example, a machine learning model may perform simple regression or multiple regression. As described above, in some implementations, the softmax function or other functions or hierarchies convert a set of real values, each associated with two or more possible classes, into a set of real values in the range (0, 1) whose sum is 1. Can be used for quashing.

In embodiments, machine learning models may perform various types of clustering. For example, a machine learning model may identify one or more previously defined clusters to which the input data is most likely to correspond. In some implementations where the machine learning model performs clustering, the machine learning model may be trained using unsupervised learning techniques.

In embodiments, a machine learning model may perform anomaly detection or outlier detection. For example, a machine learning model may identify input data that does not conform to expected patterns or other characteristics (e.g., as previously observed from previous input data). For example, anomaly detection may be used to detect fraudulent transactions or detect system failures.

In some implementations, a machine learning model may provide output data in the form of one or more recommendations. For example, machine learning models may be included in a recommendation system or engine. For example, given input data that describes previous outcomes (e.g., scores, rankings, or grades representing amounts of success or enjoyment) for a particular entity, a machine learning model can, based on the prior outcomes, produce the desired outcome. It may output suggestions or recommendations of one or more additional entities that are expected to have .

As described above, a machine learning model may be or include one or more of a variety of different types of machine learning models. Examples of these different types of machine learning models are provided below for illustration. One or more of the example models described below may be used (e.g., combined) to provide output data in response to input data. Additional models other than the example models provided below may also be used.

In some implementations, the machine learning model includes one or more classifier models, such as, for example, a linear classification model; secondary classification model; It may be or may include this. A machine learning model may include one or more regression models, such as, for example, a simple linear regression model; multiple linear regression model; Logistic regression model; stepwise regression model; Multivariate adaptive regression spline; Locally estimated scatterplot smoothing model; It may be or may include this.

In some examples, the machine learning model is one or more decision tree-based models, such as, for example, classification and/or regression trees; Chi-square automatic interaction detection decision tree; decision stump; conditional decision tree; It may be or may include this.

A machine learning model may be or include one or more kernel machines. In some implementations, a machine learning model may be or include one or more support vector machines. The machine learning model may include one or more instance-based learning models, such as, for example, a learning vector quantization model; Self-organizing map model; Locally weighted learning model; It may be or may include this. In some implementations, the machine learning model includes one or more nearest neighbor models, such as, for example, a k-nearest neighbor classification model; k-nearest neighbor regression model; It may be or may include this. The machine learning model may include, for example, one or more Bayesian models, such as a naive Bayes model; Gaussian naive Bayes model; multinomial naive Bayes model; averaged 1-dependency estimator; Bayesian network; Bayesian trust network; Hidden Markov Model; It may be or may include this.

The machine learning model may include one or more clustering models, such as, for example, a k-means clustering model; k-median clustering model; expectation maximization model; hierarchical clustering model; It may include etc.

In some implementations, the machine learning model can be modified using one or more dimensionality reduction techniques, such as, for example, principal component analysis; Kernel principal component analysis; Graph-based kernel principal component analysis; principal component regression; Partial least squares regression; Salmon Mapping; multidimensional scaling; projection pursuit; linear discriminant analysis; mixed discriminant analysis; second-order discriminant analysis; generalized discriminant analysis; flexible discriminant analysis; auto-encoding; etc. can be performed.

In some implementations, a machine learning model can be implemented using one or more reinforcement learning techniques, such as a Markov decision process; dynamic programming; Q function or Q-learning; value function approach; deep Q-network; differentiable neural computer; A3C(asynchronous advantage actor-critics); deterministic policy gradient; etc. can be performed or these can be applied.

In embodiments, artificial intelligence module 8804 may include and/or provide access to neural network module 8814. In an embodiment, neural network module 8814 is configured to train, deploy, and/or utilize an artificial neural network (or “neural networks”) on behalf of intelligent service client 8836. Note that in the description, the term machine learning model may include a neural network, and as such, neural network module 8814 may be part of machine learning module 8812. In an embodiment, neural network module 8814 may be configured to train a neural network that may be used by intelligent client 8836. Non-limiting examples of different types of neural networks include convolutional neural networks (CNNs), deep convolutional neural networks (DCNs), feedforward neural networks (including deep feedforward neural networks), and recurrent neural networks (RNNs) (including gated RNNs). (but not limited to), long-term/short-term memory (LTSM) neural networks, etc., as well as serially, in parallel, in acyclic (e.g., directed graph-based) flows, and/or intermediate decision nodes, recursive loops, etc. may include any of the neural network types described throughout this disclosure and the documents incorporated by reference herein, including but not limited to hybrids or combinations of the above, such as deployed in more complex flows that may include A neural network of a given type can take input from a data source or another neural network and provide output that is contained within the input set of the other neural network until the flow is complete and the final output is provided. In embodiments, neural network module 8814 may be utilized by other artificial intelligence modules 8804, such as machine vision module 8822, NLP module 8824, rule-based module 8828, digital twin module 8826, etc. there is. Exemplary applications of neural network module 8814 are described throughout this disclosure.

Neural networks contain groups of connected nodes, which may also be referred to as neurons or perceptrons. Neural networks can be organized into one or more layers. A neural network containing multiple layers may be referred to as a “deep” network. A deep network may include an input layer, an output layer, and one or more hidden layers located between the input layer and the output layer. Nodes in a neural network can be connected or completely disconnected.

In embodiments, the neural network may be or include one or more feedforward neural networks. In a feedforward network, the connections between nodes do not form cycles. For example, each connection may connect a node from a previous layer to a node from a later layer.

In embodiments, the neural network may be or include one or more recurrent neural networks. In some cases, at least some of the nodes of a recurrent neural network may form a cycle. Recurrent neural networks can be particularly useful for processing input data that is sequential in nature. In particular, in some cases, a recurrent neural network may transfer or retain information from a previous portion of an input data sequence to a subsequent portion of the input data sequence through the use of recursive or directed recurrent node connections.

In some examples, sequential input data may include time series data (e.g., time or images captured at different times subject to sensor data). For example, recurrent neural networks can analyze sensor data versus time to detect or predict swipe direction, perform handwriting recognition, and more. Sequential input data includes words within sentences (e.g., for natural language processing, conversation detection or processing, etc.); notes in musical composition; sequential actions taken by the user (e.g., to detect or predict sequential application usage); sequential object states; It may include etc. In some example embodiments, the recurrent neural network may be a long short-term (LSTM) recurrent neural network; gated circulation unit; Bidirectional recurrent neural network; Continuous-time recurrent neural network; neural history compressor; echo state network; Elman Network; Jordan Network; recursive neural network; Hopfield Network; fully recurrent network; Sequence-to-sequence composition; Includes etc.

In some examples, the neural network may be or include one or more acyclic sequence-to-sequence models based on self-attention, such as a transformer neural network. Details of an example transformer neural network can be found at http://papers.nips.cc/paper/7181-attention-is-all-you-need.pdf.

In embodiments, the neural network may be or include one or more convolutional neural networks. In some cases, a convolutional neural network may include one or more convolutional layers that perform convolutions on input data using learned filters. A filter may also be referred to as a kernel. Convolutional neural networks can be particularly useful for vision problems, such as when the input data includes images such as still images or video. However, convolutional neural networks can also be applied for natural language processing.

In embodiments, the neural network may be or include one or more generative networks, such as, for example, an adversarial generative neural network. Generative networks can be used to generate new data, such as new images or other content.

In an embodiment, the neural network may be or include autoencoders. In some cases, the goal of an autoencoder is to learn a representation (e.g., low-dimensional encoding) for a set of data, typically for the purpose of dimensionality reduction. For example, in some cases, an autoencoder may attempt to encode input data and then provide output data that reconstructs the input data from the encoding. Recently, the autoencoder concept has become more widely used to learn generative models of data. In some cases, autoencoders may involve additional losses beyond reconstructing the input data.

In embodiments, the neural network may be one or more other types of artificial neural networks, such as, for example, a deep Boltzmann machine; deep trust network; Stacked autoencoder; It may be or may include this. Any of the neural networks described in this application can be combined (e.g., stacked) to form more complex networks.

Figure 105 illustrates an example neural network with multiple layers. Neural network 8840 may include an input layer, a hidden layer, and an output layer, each layer including a plurality of nodes or neurons that respond to a different combination of inputs from the previous layer. Connections between neurons have numerical weights that determine how much relative influence the input has on the output value of that node. The input layer includes a plurality of input nodes 8842, 8844, 8846, which can provide information or input data (e.g., sensor data, image data, text data, audio data, etc.) from the external world to the neural network 8840. 8848 and 8850). The input data may be from different sources and may include library data (x1), simulation data (x2), user input data (x3), training data (x4) and results data (x5). Input nodes 8842, 8844, 8846, 8848, and 8850 may pass information to the next layer, and no computation may be performed by the input nodes. The hidden layer may include multiple nodes, such as nodes 8852, 8854, and 8856. Nodes within hidden layers 8852, 8854, and 8856 may process information from the input layer and pass information to the output layer based on the weights of the connections between the input layer and the hidden layer. The output layer processes information based on the weights of the connections between the hidden layer and the output layer, computes and transmits the information from the network to the outside world, such as recognizing specific objects or activities, or predicting conditions or actions. It may include an output node 8858 that is responsible for doing this.

In an embodiment, neural network 8840 may include two or more hidden layers and may be referred to as a deep neural network. The layers are structured such that a first layer detects a set of primitive patterns in input (e.g., image) data, a second layer detects patterns of patterns, and a third layer detects patterns of those patterns. In some embodiments, a node within neural network 8840 may have connections to all nodes in the immediately previous layer and the immediately next layer. Accordingly, the layer may be referred to as a fully-connected layer. In some embodiments, nodes within neural network 8840 may have connections only to some of the nodes in the immediately previous layer and the immediately next layer. Accordingly, the layer may be referred to as a sparsely-connected layer. Each neuron in a neural network consists of a weighted linear combination of its inputs, and the calculations for each neural network layer can be described as the multiplication of the input matrix and the weighting matrix. Next, a bias matrix is added to the resulting product matrix to take into account the threshold of each neuron at the next level. Additionally, an activation function is applied to each resulting value, and the resulting value is placed in the matrix for the next layer. Therefore, the output from node i in the neural network can be expressed as:

yi=f (Σxiwi+bi)

Here, f is the activation function, Σxiwi is the weighted sum of the input matrices, and bi is the bias matrix.

The activation function determines the level of activity or excitation generated in a node as a result of an input signal of a certain magnitude. The purpose of an activation function is to introduce non-linearity into the output of a neural network node, since most real-world functions are non-linear and it is desirable for neurons to be able to learn such non-linear representations. Several activation functions can be used in artificial neural networks. One example activation function is the sigmoid function σ(x), which is a continuous sigmoid monotonically increasing function that asymptotically approaches a fixed value as the input approaches plus or minus infinity. The sigmoid function σ(x) takes a real-valued input and converts it to a value between 0 and 1:

σ(x)=1/(1+exp(-x)).

Another example activation function is the tanh function, which takes a real-valued input and converts it to a value in the range [-1, 1]:

tanh(x)=2σ(2x)- 1

A third example activation function is a rectified linear unit (ReLU) function. The ReLU function takes a real-valued input and thresholds it above 0 (i.e. replaces negative values with 0):

f(x)=max(0, x).

Activation functions are provided by way of example, and in various embodiments, the neural network 8840 may be configured to include identity, binary step, logistic, soft step, tanh, arctangent, softsine, rectified linear unit (ReLU), Leaky commutation linear unit, parametric rectification. Linear units, random Rickey rectification linear units, exponential linear units, sigmoidal rectification linear activation units, adaptive piecewise linear, softplus, bent identity, softexponential, sinusoidal, sinc, Gaussian, softmax, maxout, and/ It will be clear that a variety of activation functions may be used, including but not limited to combinations of activation functions.

In the example shown in Figure 105, nodes 8842, 8844, 8846, 8848, and 8850 in the input layer can take external inputs (x1, x2, x3, x4, and x5), which can be numeric values depending on the input data set. there is. Although only five inputs are shown in Figure 105, it will be appreciated that in various implementations, a node may contain tens, hundreds, thousands, or more inputs. As discussed above, no computation is performed on the input layer, so the outputs from nodes 8842, 8844, 8846, 8848, and 8850 of the input layer are x1, x2, x3, x4, and x5, respectively, which are hidden supplied to the hierarchy. The output of node 8852 in the hidden layer may depend on the outputs from the input layers (x1, x2, x3, x4, and x5) and the weights associated with the connections (w1, w2, w3, w4, and w5). Therefore, the output from node 8852 can be calculated as:

Y ₈₈₅₂ =f(x1w1+x2w2+x3w3+x4w4+x5w5 +b ₈₈₅₂ ).

The outputs from nodes 8854 and 8856 in the hidden layer may also be computed in a similar manner and then fed to node 8858 in the output layer. Node 8858 in the output layer can perform similar calculations (using weights v1, v2, and v3 associated with the connection) as nodes 8852, 8854, and 8856 in the hidden layer:

Y ₈₈₅₈ = f(y ₈₈₅₂ v1+y ₈₈₅₄ v2+y ₈₈₅₆ v3+b ₈₈₅₈ );

Here, Y ₈₈₅₈ is the output of the neural network 8840.

As mentioned, connections between nodes within a neural network have weights associated with them, which determine how much relative influence the input value has on the output value of that node. Before the network is trained, random values are chosen for each weight. Weights are adjusted during the training process, and this adjustment of weights to determine the best set of weights that maximizes the accuracy of the neural network is referred to as training. For every input in the training data set, the output of the artificial neural network can be observed and compared to the expected output, and the error between the expected and observed outputs can be propagated back to the previous layer. The weights can be adjusted accordingly based on the error. This process is repeated until the output error is below a predetermined threshold.

In embodiments, backpropagation (e.g., backward propagation of errors) is used in conjunction with optimization methods such as gradient descent to adjust weights and update neural network properties. Backpropagation can be a supervised training technique that learns from labeled training data and errors at nodes by changing the parameters of the neural network to reduce errors. For example, the results of forward propagation (e.g., output activation value(s)) determined using training input data are compared to corresponding known reference output data to calculate the loss function gradient. The gradient can then be used in an optimization method to determine new updated weights in an attempt to minimize the loss function. For example, to measure error, the mean square error is determined using the equation:

E=(target-output) ²

To determine the gradient for weight “w”, the partial derivative of the error with respect to the weight can be determined, where:

Gradient=∂E/∂w

Computation of the partial derivative of the error with respect to the weights can flow backwards through the node levels of the neural network. Then, a portion of the gradient (e.g., ratio, percentage, etc.) is subtracted from the weight to determine the updated weight. That part can be designated as the learning rate "a". Accordingly, an exemplary equation for determining the updated weights is given by the formula:

w new =w old -α∂E/∂w

Choose the learning rate so that it is not too small (e.g., a rate that is too small may lead to slow convergence to the desired weights) and not too large (e.g., a rate that is too large may cause the weights to not converge to the desired weights). It has to be.

After adjusting the weights, the network should perform better than before for the same input because the weights have now been adjusted to minimize errors.

As mentioned, neural networks may include convolutional neural networks (CNNs). CNN is a specialized neural network for processing data with a known grid-like topology, such as image data. Therefore, CNNs are commonly used in classification, object recognition, and computer vision applications, but they can also be used for other types of pattern recognition, such as conversation and language processing.

Convolutional neural networks learn highly nonlinear mappings by interconnecting layers of artificial neurons arranged in many different layers with activation functions that form layer dependencies. It includes one or more convolutional layers interspersed with one or more subsampling layers and non-linear layers, which are typically followed by one or more fully connected layers.

106, CNN 8860 includes an input layer with input images 8862 to be classified by CNN 8860, interspersed with one or more activation or non-linear layers (e.g., ReLU), and a pooling or subsampling layer. a hidden layer, which in turn contains one or more convolutional layers, and an output layer, which typically contains one or more fully connected layers. The input image 8862 may be represented by a matrix of pixels and may have multiple channels. For example, a color image may have red, green, and blue channels, respectively representing the red, green, and blue (RGB) components of the input image. Each channel can be represented by a 2-D matrix of pixels with pixel values ranging from 0 to 255. Meanwhile, a gray scale image can have only one channel. The next section describes the processing of a single image channel using CNN 8860. It will be appreciated that multiple channels may be handled in a similar manner.

As shown, input image 8862 may be processed by a hidden layer that includes a set of convolutional and activation layers 8864 and 8868, respectively, followed by pooling layers 8866 and 8870.

The convolutional layer of a convolutional neural network serves as a feature extractor that can learn the input image and decompose it into hierarchical features. A convolution layer can perform a convolution operation on an input image, where a filter (also called a kernel or feature detector) can slide over the input image with a specific step size (also called a stride). For every position (or step), an element-wise multiplication between the filter matrix and the nested matrix within the input image is computed and summed to obtain a final value that represents a single element of the output matrix that makes up the feature map. A feature map refers to image data representing various features of input image data, and may have smaller dimensions compared to the input image. The activation or non-linear layers use different non-linear trigger functions for signaling distinct identifications of likely features on each hidden layer. The nonlinear layer uses a variety of specific functions to implement nonlinear triggering, including rectified linear units (ReLU), hyperbolic tangent, and absolute and sigmoid functions of hyperbolic tangent. In one implementation, ReLU activation implements the function y=max(x, 0) and keeps the input and output sizes of the layer the same. The advantage of using ReLU is that convolutional neural networks are trained many times faster. ReLU is a discontinuous, non-saturating activation function that is linear with respect to the input if the input value is greater than 0 and is 0 otherwise.

As shown in Figure 106, the first convolution and activation layer 8864 consists of a number of output matrices (or feature maps) 8872 followed by a non-linear operation (e.g., ReLU). Convolution can be performed on the input image 8862 using a filter. The number of filters used may be referred to as the depth of the convolutional layer. Accordingly, the first convolution and activation layer 8864 in the example of Figure 106 has a depth of 3 and uses 3 filters to generate 3 feature maps. Feature map 8872 can then be passed to a first pooling layer that can subsample or downsample the feature map using a pooling function to generate output matrix 8874. The pooling function replaces feature maps with summary statistics to reduce the spatial dimension of the extracted feature maps, thereby reducing the number of parameters and computations in the network. Therefore, the pooling layer reduces the dimensionality of the feature map while retaining the most important information. Pooling functions can also be used to introduce translation invariance into a neural network, such that small transformations on the input do not change the pooled output. Different pooling functions can be used in the pooling layer, including max pooling, average pooling, and 12-norm pooling.

The output matrix 8874 is then processed by a second convolution and activation layer 8868 to perform convolution and non-linear activation operations (e.g., ReLU) as described above to produce a feature map 8876. can be created. In the example shown in Figure 106, the second convolution and activation layer 8868 may have a depth of 5. Feature map 8876 can then be passed to a pooling layer 8870, where feature map 8876 can be subsampled or downsampled to generate output matrix 8878.

The output matrix 8878 produced by the pooling layer 8870 is then processed by one or more fully connected layers 8880 that form part of the output layer of the CNN 8860. The fully connected layer 8880 has full connectivity with all feature maps in the output matrix 8878 of the pooling layer 8870. In an embodiment, the fully connected layer 8880 takes the output matrix 8878 generated by the pooling layer 8870 as input in the form of a vector and outputs a feature vector containing information of the structure within the input image. A level decision can be made. In an embodiment, fully connected layer 8880 may use a softmax function to classify objects in input image 8862 into one of several categories. The softmax function can be used as an activation function in the output layer, taking a vector of real-valued scores and mapping it to a vector of values between 0 and 1, which sum to 1. In embodiments, other classifiers may be used, such as a support vector machine (SVM) classifier.

In embodiments, one or more normalization layers may be added to CNN 8860 to normalize the output of the convolutional filter. Regularization layers can provide whitening or lateral suppression, avoid vanishing or exploding gradients, stabilize training, and enable learning at higher rates and faster convergence. In an embodiment, a normalization layer is added after the convolution layer, but before the activation layer.

Accordingly, CNN 8860 can be viewed as a multiple set of convolutional, activation, pooling, normalization, and fully connected layers stacked together to learn, enhance, and extract intrinsic features and patterns in the input image 8862. As used in this application, a layer has similar functions by mathematical or other functional means to process the received input to generate/derive an output for the next layer to one or more other components for further processing within the CNN 8860. It can refer to one or more components that operate as .

An initial layer of CNN 8860, such as a convolutional layer, may extract low-level features, such as edges and/or gradients, from the input image 8862. Subsequent layers can extract or detect progressively more complex features and patterns, such as the presence of curvature and texture in the image data, etc. The output of each layer can serve as an input to a subsequent layer in CNN 8860 to learn a hierarchical feature representation from data in the input image 8862. This allows convolutional neural networks to efficiently learn increasingly complex and abstract visual concepts.

Although only two convolutional layers are shown in the example, the present disclosure is not limited to the example architecture, and the CNN (8860) architecture may have any number of layers in total, and any number for convolution, activation, and pooling. It may include a hierarchy of For example, there have been many variations and improvements over the basic CNN model described above. Some examples are Alexnet, GoogLeNet, VGGNet (stacks many layers including a narrow convolutional layer followed by a max-pooling layer), residual neural network or ResNet (uses residual blocks and skip connections to learn residual mappings), DenseNet (connects each layer of the CNN to every other layer in a feed-forward manner), Squeeze and excitation network (includes global context in features), and AmobeaNet (optimal architecture for image recognition) uses evolutionary algorithms to search and find.

TRAINING OF CONVOLUTIONAL NEURAL NETWORK

The training process of a convolutional neural network, such as CNN 8860, may be similar to the training process discussed in FIG. 105 with respect to neural network 8840.

In an embodiment, all parameters and weights (including weights in filters and weights for fully connected layers) are initially assigned (e.g., randomly assigned). Then, during training, the training image or images in which objects are detected and classified are provided as input to CNN 8860, and the CNN performs a forward propagation step. That is, the CNN 8860 applies convolution, nonlinear activation, and pooling layers to each training image to determine a classification vector (i.e., detects and classifies each training image). These classification vectors are compared to predetermined classification vectors. The error (e.g., sum of squares of differences, log loss, softmax log loss) between the CNN's classification vector and a predetermined classification vector is determined. These errors are then used to update the CNN's weights and parameters in a backpropagation process, which may use gradient descent and may include one or more iterations. The training process is repeated for each training image in the training set.

The training process and inference process described above may be performed on hardware, software, or a combination of hardware and software. However, training a convolutional neural network such as CNN 8860 or using a trained CNN for inference typically requires a significant amount of computational power, for example, to perform matrix multiplication or convolution. Therefore, specialized hardware circuits such as graphics processing units (GPUs), tensor processing units (TPUs), neural network processing units (NPUs), FPGAs, ASICs, or other highly parallel processing circuits may be used for training and/or inference. there is. Training and inference can be performed in the cloud, in a data center, or on the device.

REGION BASED CNN (RCNN) AND OBJECT DETECTION

In embodiments, an object detection model extends the functionality of a CNN-based image classification neural network model by not only classifying an object but also determining its location in the image in terms of its bounding box. Region-based CNN (R-CNN) method is used to extract regions of interest (ROI), where each ROI is a rectangle that can represent the boundaries of objects within the image. Conceptually, R-CNN operates in two steps. In the first step, the region proposal method generates all potential bounding box candidates in the image. In the second stage, for all proposals, a CNN classifier is applied to distinguish objects. Alternatively, a fast R-CNN architecture can be used that integrates feature extractors and classifiers into a unified network. Other faster R-CNNs can be used, which integrate Region Proposal Network (RPN) and fast R-CNN into an end-to-end trainable framework. Mask R-CNN adds instance segmentation, while Mesh R-CNN adds the ability to generate 3D meshes from 2D images.

In embodiments, artificial intelligence module 8804 may provide access to and/or integrate with robotic process automation (RPA) module 8816. RPA module 8816 may facilitate computer automation, among other things, creating and validating workflows. In embodiments, RPA module 8816 may monitor human interactions with various systems to learn patterns and processes performed by humans in the performance of each task. This may include observation of human actions involving interactions with hardware elements, software interfaces, and other elements. Observations include, among many other examples, field observations of humans performing actual tasks, as well as simulations or other instances of humans performing actions with the explicit intent of providing a training data set or input for the RPA system. It may involve observation of activity, for example in this case a human tagging or labeling a training data set with features that assist the RPA system in learning to recognize or classify features or objects. In embodiments, RPA module 8816 may learn to perform specific tasks based on learned patterns and processes, such that tasks may be performed by RPA module 8816 instead of or in support of a human decision maker. . Examples of RPA modules 8816 may include those in this disclosure and documents incorporated by reference herein, and may involve automation of any of the broad value chain network activities or entities described therein.

In embodiments, artificial intelligence module 8804 may include and/or provide access to analytics module 8818. In embodiments, analysis module 8818 is configured to perform various analysis processes on data output from value chain entities or other data sources. In example embodiments, analytics information generated by analytics module 8818 may facilitate quantification of system performance compared to a set of goals and/or metrics. Goals and/or metrics may be pre-configured, dynamically determined from operational results, etc. Examples of analysis processes that may be performed by analysis module 8818 are discussed below and in documents incorporated by reference into this application. In some example implementations, the analytics process includes (among many others) goals that involve coordination of value chain activities and demand intelligence, such as entailing anticipating demand for a set of related items by location and time; and /or may involve tracking specific metrics.

In embodiments, artificial intelligence module 8804 may include and/or provide access to digital twin module 8820. Digital twin module 8820 may include any of the wide range of features and capabilities described in this application. In embodiments, the digital twin module 8820 may be used to connect different types of digital twins, such as physical environment twins, robotic operational unit twins, logistics twins, executive digital twins, organizational digital twins, role-based digital twins, etc., among others. It can be configured to provide an execution environment for In embodiments, digital twin module 8820 may be configured according to digital twin systems and/or modules described elsewhere throughout this disclosure. In an example embodiment, digital twin module 8820 may be configured to create a digital twin requested by intelligence client 8836. Additionally, the digital twin module 8820 may be configured with an interface such as an API for receiving information from an external data source. For example, digital twin module 8820 may receive real-time data from sensor systems in a machine, vehicle, robot, or other device, and/or sensor systems in the physical environment in which the device operates. In embodiments, digital twin module 8820 may receive digital twin data from other suitable data sources, such as third party services (e.g., weather services, traffic data services, logistics systems and databases, etc.). In an embodiment, the digital twin module 8820 may support value chain network entities such as supply chain infrastructure entities, transportation or logistics entities, containers, goods, etc., as well as demand networks such as customers, merchants, stores, points of sale, points of use, etc. It may contain digital twin data representing the characteristics, status, etc. of the entity. The digital twin module 8820 is or is integrated with an interface (e.g., a control tower or dashboard) for coordination of supply and demand, including coordination of automation within supply chain activities and demand management activities. , can be linked to it, or interact with it in some other way.

In embodiments, digital twin module 8820 may provide access to and manage a library of digital twins. Artificial intelligence module 8804 may access the library to perform functions, such as simulation of actions in a given environment in response to specific stimuli.

In embodiments, artificial intelligence module 8804 may include and/or provide access to machine vision module 8822. In an embodiment, machine vision module 8822 is configured to process images (e.g., captured by a camera) to detect and classify objects within the image. In an embodiment, machine vision module 8822 receives one or more images (which may be single still shot images or frames of a video feed) and detects a “blob” in the image (e.g., using edge detection techniques, etc.). )". Machine vision module 8822 may then classify the blob. In some embodiments, machine vision module 8822 utilizes one or more machine learned image classification models and/or neural networks (e.g., convolutional neural networks) to classify blobs within an image. In some embodiments, machine vision module 8822 may perform feature extraction on the image and/or each blob within the image prior to classification. In some embodiments, machine vision module 8822 may utilize classifications made from previous images to confirm or update the classification(s) from previous images. For example, if an object detected in a previous frame was classified with a lower confidence score (e.g., the object is partially occluded or out of focus), machine vision module 8822 may verify the classification or If module 8822 is able to determine the object's classification with higher confidence, it may update. In an embodiment, machine vision module 8822 is configured to detect occlusion, such as an object that may be occluded by another object. In embodiments, machine vision module 8822 may provide additional input to assist with the image classification task, such as from radar, sonar, a digital twin of the environment (which may show the location of known objects), and/or the like. Receive. In some embodiments, machine vision module 8822 may include or interface with a liquid lens. In these embodiments, liquid lenses facilitate improved machine vision (e.g., when focusing at multiple distances is required by the robot's environment and tasks) and/or other machine vision tasks enabled by the liquid lenses. You can do it.

In embodiments, artificial intelligence module 8804 may include and/or provide access to a natural language processing (NLP) module 8824. In an embodiment, NLP module 8824 performs natural language tasks on behalf of intelligent service client 8836. Examples of natural language processing technologies include conversation recognition, conversation segmentation, speaker diarization, text-to-speech, lemmatization, morphological segmentation, and parts-of-speech tagging. , stemming, syntactic analysis, lexical analysis, etc., but is not limited to these. In embodiments, NLP module 8824 may enable voice commands received from a human. In an embodiment, NLP module 8824 may receive an audio stream (e.g., from a microphone) and perform speech-to-text conversion on the audio stream to obtain a transcription of the audio stream. NLP module 8824 may process text (e.g., a transcription of an audio stream) to determine the meaning of the text using various NLP techniques (e.g., NLP models, neural networks, etc.). In embodiments, NLP module 8824 may determine an action or command spoken in the audio stream based on the results of the NLP. In an embodiment, NLP module 8824 may output the results of NLP to intelligent service client 8836.

In an embodiment, NLP module 8824 provides intelligent service client 8836 with the ability to communicate with a human user as well as parse one or more conversational voice commands provided by the human user to perform one or more tasks. The NLP module 8824 can perform conversation recognition to recognize voice commands, natural language understanding to parse and derive meaning from commands, and natural language generation to generate voice responses to the user when processing user commands. . In some embodiments, NLP module 8824 enables intelligent service client 8836 to understand commands and provide a response to the user upon successful completion of the task by intelligent service client 8836. In embodiments, NLP module 8824 may formulate and ask questions to the user when the context of the user's request is not entirely clear. In embodiments, NLP module 8824 may use input received from one or more sensors, including vision sensors, location-based data (e.g., GPS data), to determine contextual information associated with the processed conversation or text data. Available.

In an embodiment, the NLP module 8824 uses neural networks such as recurrent neural networks, long short-term memory (LSTM), gated recurrent units (GRU), transformer neural networks, convolutional neural networks, etc. when performing NLP tasks.

107 illustrates an example neural network 8800 for implementing NLP module 8824. In the illustrated example, the exemplary neural network is a transformer neural network. In this example, transformer neural network 8800 includes three input stages and five output stages to transform the input sequence into an output sequence. Exemplary transformers include encoder 8802 and decoder 8804. Encoder 8802 processes input and decoder 8804 generates output probabilities, for example. Encoder 8802 includes three stages and decoder 8804 includes five stages. Encoder 8802 Stage 1 represents the input as a sequence of positional encodings added to the embedded input. Encoder 8802 stage 2 and stage 3 include N layers (e.g., N=6, etc.), where each layer includes a location-specific feedforward neural network (FNN) and an attention-based sublayer. Each attention-based sublayer in Encoder 8802 Stage 2 includes four linear projections and multi-head attention logic to be added and normalized to be fed to the location-specific FNN in Encoder 8802 Stage 3. Encoder 8802 stages 2 and 3 utilize residual concatenation followed by a normalization layer at the output.

The example decoder 8804 processes the output embedding as its input to help ensure that the prediction for position i depends on positions before/below i, with the output embedding shifted right by one position. In stage 2 of decoder 8804, the masked multi-head attention is modified to prevent positions from attending to subsequent positions. Stages 3-4 of the decoder 8804 include N layers (e.g., N=6, etc.), where each layer includes a location-specific FNN and two attention-based sublayers. Each attention-based sublayer in decoder 8804 stage 3 includes four linear projections and multi-head attention logic to be added and normalized to be provided to the location-specific FNN in decoder 8804 stage 4. Decoder 8804 stages 2-4 use residual concatenation followed by a normalization layer at the output. Decoder 8804 stage 5 provides a linear transformation followed by a softmax function to normalize the resulting vector of K numbers into a probability distribution 8806 containing K probabilities proportional to the exponents of the K input numbers. .

Additional examples of neural networks can be found elsewhere in this disclosure (e.g., Figures 78-103).

Referring back to FIG. 104 , in embodiments, the artificial intelligence module 8804 also includes and/or provides access to a rules-based module 8828 that may be integrated into or accessed by the intelligent services client 8836. can do. In some embodiments, rule-based module 8828 may consist of programmatic logic that defines sets of rules and other conditions that trigger specific actions that may be performed in conjunction with an intelligent client. In embodiments, rule-based module 8828 may consist of programmatic logic that receives input and determines whether one or more rules are satisfied based on the input. If the condition is met, the rule-based module 8828 determines an action to perform, which can be output to the request intelligence service client 8836. Data received by the rule-based engine may be received from an intelligence service input source 8832 and/or within an artificial intelligence module 8804, such as a machine vision module 8822, a neural network module 8814, an ML module 8812, etc. May be requested from other modules. For example, rule-based module 8828 may classify objects within the field of view of a mobile system (e.g., robots, autonomous vehicles, etc.) from a machine vision system and/or receive sensor data from a LiDAR sensor of the mobile system. and, in response, determine whether the mobile system should continue its path, change its course, or stop. In embodiments, rule-based module 8828 may be configured to make other appropriate rule-based decisions on behalf of each client 8836, examples of which are discussed throughout this disclosure. In some embodiments, the rules-based engine may apply governance standards and/or analysis modules, which are described in more detail below.

In an embodiment, the artificial intelligence module 8804 interfaces with, and in response to, an intelligent service controller 8802 that is configured to determine the type of request issued by the intelligent service client 8836, such that the artificial intelligence module 8804 uses artificial intelligence when responding to the request. Module 8804 may determine the set of governance standards and/or analyzes to be applied. In an embodiment, intelligent service controller 8802 may include an analytics management module 8806, a set of analytics modules 8808, and a governance library 8810.

In an embodiment, the intelligent service controller 8802 is configured to determine the type of request issued by the intelligent service client 8836 and, in response, determine the governance standards to be applied by the artificial intelligence module 8804 when responding to the request. and/or determine a set of analyses. In an embodiment, intelligent service controller 8802 may include an analytics management module 8806, a set of analytics modules 8808, and a governance library 8810. In an embodiment, analytics management module 8806 receives artificial intelligence module 8804 requests and determines governance standards and/or analytics associated with the requests. In embodiments, analysis management module 8806 may determine governance standards to apply to a request based on the type of decision requested and/or whether a particular analysis should be performed for the requested decision. For example, a request for a control decision that causes the intelligent service client 8836 to perform an action may be tied to a specific set of applicable governance standards, such as safety standards, legal standards, quality standards, etc., and/or One or more analyzes may be coupled to control decisions, such as analysis, safety analysis, engineering analysis, etc.

In some embodiments, analytics management module 8806 may determine governance standards to apply to a decision request based on one or more conditions. Non-limiting examples of such conditions may include the type of decision being requested, the geographic location in which the decision is being made, the environment the decision will affect, current or predicted environmental conditions of the environment, etc. In embodiments, governance standards may be defined as a set of standards libraries stored in governance library 8810. In embodiments, a standard library may define conditions, thresholds, rules, recommendations, or other suitable parameters under which decisions may be analyzed. Examples of standard libraries may include legal standard libraries, regulatory standard libraries, quality standard libraries, engineering standard libraries, safety standard libraries, financial standard libraries, and/or other suitable types of standard libraries. In an embodiment, governance library 8810 may include an index that indexes specific standards defined in each standard library based on different conditions. Examples of conditions may be jurisdictions or geographic areas to which a particular standard applies, environmental conditions to which a particular standard applies, device types to which a particular standard applies, materials or products to which a particular standard applies, etc.

In some embodiments, the analytics management module 8806 may determine an appropriate set of standards that should be applied in connection with a particular decision, and the artificial intelligence module 8804 may utilize the associated governance standards in determining the decision. Module 8804 may be provided with an appropriate set of standards. In such embodiments, artificial intelligence module 8804 may be configured to apply standards in the decision-making process such that decisions output by artificial intelligence module 8804 are consistent with associated governance standards. It is understood that the standard library within the governance library may be defined by the platform provider, customer, and/or third party. Standards may be government standards, industry standards, customer standards, or other suitable sources. In embodiments, each set of standards may include a set of conditions that associate each set of standards such that the conditions can be used to determine which standard to apply in a given situation.

In some embodiments, analysis management module 8806 may determine one or more analyzes to be performed for a particular decision and may provide corresponding analysis modules 8808 to artificial intelligence module 8804 to perform these analyses, , the artificial intelligence module 8804 utilizes the corresponding analysis module 8808 to analyze the decision before outputting the decision to the requesting client. In embodiments, analysis module 8808 may include modules configured to perform a particular analysis on a particular type of decision, where each module is configured by a processing system hosting an instance of intelligence service 8800. It runs. Non-limiting examples of analysis modules 8808 include risk analysis module(s), security analysis module(s), decision tree analysis module(s), ethics analysis module(s), failure mode and effects (FMEA) analysis module ( s), hazard analysis module(s), quality analysis module(s), safety analysis module(s), regulatory analysis module(s), legal analysis module(s), and/or other suitable analysis module(s). It can be included.

In some embodiments, analysis management module 8806 is configured to determine which type of analysis to perform based on the type of decision requested by intelligent service client 8836. In some of these embodiments, analysis management module 8806 may include an index or other suitable mechanism to identify a set of analysis modules 8808 based on the requested decision type. In this embodiment, analysis management module 8806 can receive a decision type and determine a set of analysis modules 8808 to run based on the decision type. Additionally or alternatively, one or more governance standards may define when certain analyzes are performed. For example, engineering standards may define which scenarios require FMEA analysis. In this example, an engineering standard may be tied to a request for a specific type of decision, and the engineering standard may define a scenario in which the FMEA analysis is performed. In this example, artificial intelligence module 8804 may execute a safety analysis module and/or a risk analysis module and determine alternative decisions about whether an action would violate a legal or safety standard. In response to analyzing the proposed decision, artificial intelligence module 8804 may optionally output proposed conditions based on the results of the performed analysis. If the decision is accepted, the artificial intelligence module 8804 may output the decision to the requesting intelligence service client 8836. If the proposed configuration is flagged by one or more of the analyzes, artificial intelligence module 8804 may determine an alternative decision and run the analysis on the alternative proposed decision until a matching decision is obtained.

Note here that, in some embodiments, one or more analysis modules 8808 may themselves be defined as standards, and one or more associated standards used together may include specific analyzes. For example, applicable safety standards may require a hazard analysis or more acceptable method to be used. In this example, ISO standards for overall processes and documentation, and ASTM standards for narrowly defined procedures may be used to complete the risk analysis required by the safety governance standard.

As mentioned, the above-described framework of intelligent services 8800 may be applied and/or utilized by various entities in the value chain. For example, in some embodiments, a platform-level intelligence system may be configured with the full capabilities of intelligence services 8800, and specific configurations of intelligence services 8800 may be provisioned for each value chain entity. Additionally, in some embodiments, the intelligent service client 8836 may transfer intelligent system tasks to a higher level value chain entity (e.g., edge-level or platform-level) when the intelligent service client 8836 is unable to perform the task autonomously. level). Note that in some embodiments, intelligent service controller 8802 may direct intelligent tasks to lower-level components. Additionally, in some implementations, intelligence service 8800 may be configured to output a default action when a decision cannot be reached by intelligence service 8800 and/or a higher or lower level intelligence system. In some of these implementations, default decisions may be defined in rules and/or standard libraries.

REINFORCEMENT LEARNING TO DETERMINE OPTIMAL POLICY

Reinforcement learning (RL: reinforcement learning) is a machine learning technology in which an agent repeatedly learns the optimal policy through interaction with the environment. In RL, the agent must discover the correct action by trial and error to maximize some notion of long-term reward. Specifically, in a system using RL, there are two entities: (1) the environment and (2) the agent. An agent is a computer program component that connects to the environment so that it can sense the state of the environment as well as perform actions on the environment. At each step of interaction, the agent senses the current state s of the environment and selects an action a to take. An action changes the state of the environment, and the value of this state transition is communicated to the agent by a reward signal r, where the magnitude of r indicates the desirability of the action. Over time, the agent builds a policy π that specifies the action the agent will take for each state of the environment.

Formally, in reinforcement learning, a discrete set S of environmental states; A discrete set of agent actions; and there is a set R of scalar enhancement signals. After learning, the system generates a policy π that defines the value of taking action aεA in state sεS. The policy defines Qπ(s, a) as the expected return value starting from s, taking action a, and following policy π.

Reinforcement learning agents are trained on a policy through repeated exposure to various states, allow the agent to select actions according to the policy, and provide rewards based on a function designed to reward desirable behavior. Based on the reward feedback, the system can “learn” the policy and is trained to produce desirable actions. For example, for a navigation policy, an RL agent iteratively evaluates its state (e.g., location, distance from the target object), selects an action (e.g., uses a motor to move toward the target object), and an action may be evaluated using a reward signal that provides an indication of the success of the action (e.g., a reward of +10 if the movement reduces the distance between the mobile system and the target object, and -10 reward for increasing this distance). Similarly, the RL agent repeatedly acquires an image of the target object to be grasped, attempts to grasp the object, evaluates the attempt, and then evaluates the attempts of the preceding iteration(s) to assist in deciding the next attempt. It can be trained on a phage policy by executing subsequent iterations using

There can be several approaches to train RL agents in policy. Imitation learning is a core approach where agents learn from state/action pairs, where the actions are those selected by an expert (e.g., a human) in response to the observed state. Imitation learning not only solves sample inefficiency or computability problems, but also makes the training process safer. A RL agent can derive multiple examples of state/action pairs by observing a human (e.g., navigating toward and grasping a target object) and uses these as a basis for training policies. Behavior replication (BC), which focuses on learning experts' policies using supervised learning, is an example of an imitation learning approach.

Value-based learning approaches aim to find a policy containing a sequence of actions that maximizes the expected value of future rewards (or minimizes the expected cost). The RL agent can learn the value/cost function and then derive a policy for it. Two different expected values are often mentioned: state value (V(s)) and action value (Q(s,a)), respectively. The state value function V(s) represents the value associated with the agent in each state, while the action value function Q(s,a) represents the value associated with the agent in state s and performing action a. Value-based learning approaches work by approximating the optimal value (V ^* or Q ^* ) and then deriving the optimal policy. For example, the optimal value function Q ^* (s, a) can be identified by finding the sequence of actions that maximizes the state-action value function Q(s, a). The optimal policy for each state can be derived by identifying the highest action that can be taken from each state.

π ^* (s)=argmax Q ^* (s,a)

The Bellman optimality equation can be applied to iteratively compute the value function as the actions in the sequence are executed and the mobile system transitions from one state to another. The optimal value function Q ^* (s,a) follows the Bellman optimality equation and can be expressed as:

Q ^* (s _t , a _t ) = E [r _t+1 +γmax Q ^* (s _t+1 ,a _t+1 )]

Policy-based learning approaches directly optimize the policy function π using suitable optimization techniques (e.g., stochastic gradient descent) to fine-tune the vector of parameters without computing the value function. Policy-based learning approaches are typically effective in high-dimensional or continuous action spaces.

Figure 108 illustrates an approach that is based on reinforcement learning and includes evaluation of various states, actions, and rewards in determining the optimal policy for executing one or more tasks by a mobile system.

At 8902, a reinforcement learning agent (e.g., of intelligent service system 8900) receives sensor information including a plurality of images captured by a mobile system in the environment. Analysis of one or more of these images may enable the agent to determine a first state associated with the mobile system at 8904. Data representing the first state includes information about the environment, such as images, sounds, temperature, or time, and information about the mobile system, including the location, speed, internal state of the mobile system (e.g., battery life, clock settings), etc. may include.

At 8906, 8908, and 8910, various potential actions in response to the state may be determined. Some examples of potential actions include providing control commands to actuators, motors, wheels, wing flaps, or other components that control the agent's speed, acceleration, orientation, or position; Changing the agent's internal settings, such as placing certain components into sleep mode to preserve battery life; changing direction if the agent is at risk of colliding with an obstacle object; Obtaining or transmitting data; This includes attempting to grab the target object, etc.

At 8912, 8914 and 8916, an expected reward may be determined for each potential action based on a reward function. For each potential action determined, an expected reward may be determined based on the reward function. Rewards may be based on desired outcomes such as avoiding obstacles, conserving power, or acquiring data. If the action produces a desired outcome (e.g., avoiding an obstacle), the reward is high; Otherwise, rewards may be low.

Agents can also look into the future to analyze whether there may be opportunities to realize higher rewards in the future. At 8918, 8920, and 8922, the agent may determine future states resulting from potential actions at 8906, 8908, and 8910, respectively.

For each of the future states predicted at 8918, 8920, and 8922, one or more future actions may be determined and evaluated. In steps 8924, 8926, and 8928, for example, a value or other indicator of expected reward associated with one or more of the future actions may be developed. The expected reward associated with one or more future actions can be evaluated by comparing the values of the reward function associated with each future action.

At 8930, an action may be selected based on a comparison of expected current and future rewards.

In embodiments, a reinforcement learning agent may be pre-trained through simulation in a digital twin system. In embodiments, reinforcement agents may be pre-trained using behavior replication. In an embodiment, the reinforcement agent is a deep Q-network (DQN), a double deep Q-network (DDQN), a deep deterministic policy gradient (DDPG), a soft actor critic (SAC), and an advantage actor critic. It can be trained using a deep reinforcement learning algorithm selected from actor critic (A2C), asynchronous advantage actor critic (A3C), proximity policy optimization (PPG), and trust region policy optimization (TRPO).

In embodiments, a reinforcement learning agent may balance exploitation (of current knowledge) and exploration (of unknown territory) while traversing the action space. For example, an agent may follow an ε-greedy policy by taking the optimal action most of the time with probability 1-ε while randomly choosing exploration occasionally with probability ε, where ε is 0<ε. This is a parameter that satisfies <1.

SPECIALIZED CHIPS

109-113 provide various system functionality for use in various situations, and illustrate a plurality of specialized chips that may be utilized in systems described herein and/or that may provide functionality described herein. do. As described in more detail below, chip functionality is configurable for specific situations and to solve specific tasks. Accordingly, using the functionality of one or more of the chips, systems of systems such as those described in this application can be more easily created, configured, deployed, and reconfigured. Any of the chips may be used in the various systems described herein and by various value chain entities in ways that will be apparent from the disclosure of the capabilities of each chip.

109 illustrates physical orientation determination chips 9100, one or more of which may be used to determine data regarding one or more physical orientations as described herein. Chip 9100 may be used by any value chain entity utilizing a mobile system. In embodiments, chip(s) 9100 may use artificial intelligence (AI) and other technologies to determine the physical orientation of the mobile system. As described herein, chip(s) 9100 may receive one or more inputs 9192 from a mobile system and perform one or more AI-assisted functions to determine the physical orientation of the mobile system. Chip(s) 9100 may then transmit an output 9194 indicating the determined physical orientation. Chip(s) 9100 may be part of a mobile system (e.g., a robot) and/or may be part of a different device that receives input 9192 from the mobile system (e.g., a base station that communicates with the robot). there is. A mobile system may include any system that is mobile and/or has one or more mobile components as described herein.

The physical orientation(s) determined by chip(s) 9100 may be any real reference point/frame (e.g., within the solar system, GPS coordinates, coordinates within another system, etc.) or a simulated reference point/frame (e.g., environment coordinates with a digital twin or other virtual space). In embodiments, physical orientation can be location, rotation/orientation (e.g., the direction the mobile system is facing and/or the angle at which the mobile system is rotated), and tilt (e.g., the amount the mobile system is tilted in one or more directions). , velocity, and/or acceleration, each of which can be for any real or simulated point/frame. Accordingly, output(s) 9194 may include one or more data structures representing various orientation information.

In embodiments, chip(s) 9100 may determine and/or output the orientation of the entire mobile system. Additionally or alternatively, chip(s) 9100 may determine and/or output the orientation of one or more components (e.g., rims, wheels, mechanisms, attachments, or other components) of the mobile system.

In embodiments, chip(s) 9100 may be modular component(s) that can be integrated with mobile systems in a variety of ways. As discussed above, the chip(s) may be integrated with a mobile system and/or with a system that communicates with the mobile system. To facilitate such modularity, chip(s) 9100 may be provided partially or fully within a housing (not shown) and may include electrical connectors, optical connectors, and/or wireless connectors (e.g., antennas, may receive an input 9192 and/or provide an output 9194 via an induction coil, etc.). Additionally or alternatively, chip(s) 9100 may be integrated with other circuits, processors, systems, etc. on one or multiple substrates/chips.

Chip(s) 9100 may include one or more system-on-chip (SOC), integrated circuit (IC), application specific integrated circuit (ASIC) components to provide the functionality and/or any other functionality ascribed to chip 9100. It may be and/or may include the like. For example, chip 9100 may be provided as part of a SOC that also provides other functionality described in this application. Generally, the components of chip 9100 may include one or more general purpose processing chips configured using software instructions or other code and/or special purpose processing chips (e.g. For example, ASIC) may be included.

Multiple chip(s) 9100 may be used to perform the functions described in this application. For example, multiple chip(s) 9100 may be used to determine physical orientation data more quickly, offloading more complex calculations from one chip 9100 to another chip 9100 with a better power source. Thus, serial, parallel, and/or other processing techniques may be used to more efficiently determine physical orientation data. As another example, one chip 9100 may be used to provide physical orientation data for one component of a mobile system (e.g., left arm/leg/wheel) while the other chip 9100 may be used to provide physical orientation data for one component of a mobile system (e.g., left arm/leg/wheel) It may be used to provide physical orientation data for a second component of the system (eg, right arm/leg/wheel).

In an embodiment, physical input interface 9102 receives one or more inputs 9192 to physical orientation determination chip 9100 as described herein. Input 9192 may be transmitted to physical input interface 9102 by other chips, circuits, modules, and/or other components of the mobile system. For example, input data may come from a sensor, sensor processing chip/module/circuit, antenna, storage device, network interface, or any other data source for chip(s) 9100 as described herein. You can. Physical input interface 9102 may be coupled to the source(s) of input 9192 via a wired or wireless connection. Input 9192 may include one or more of position signals/data, accelerometer, gyroscope, or other relative motion data, images, video, or other vision data, as well as LIDAR data, radar data, sonar data, etc. . Input 9192 may also include data that may be stored in storage 9150, such as images for image library 9152, data for environmental digital twin 9154 (e.g., a digital representation of the environment surrounding the mobile system) ), one or more system feature(s) 9156, and/or one or more intelligence module(s) 9158.

As noted above, the output data 9194 transmitted from the physical output interface 9104 may include data indicative of position, rotation/direction, tilt, velocity, and/or acceleration as determined by the chip 9100. It may contain more than one. In embodiments, the output of chip 9100 may be transmitted by physical output interface 9104 to other chips, circuits, modules, and/or other components as described herein. Physical output interface 9104 may be connected to these components via wired or wireless connections.

In an embodiment, chip 9100 may include one or more of location module 9110, relative motion module 9120, machine vision module 9130, and orientation module 9140. In an embodiment, location module 9110 may include circuitry 9112-9116 to determine and output a location (e.g., GPS coordinates) based on input 9192. Additionally or alternatively, chip 9100 may include circuitry 9122 for determining and outputting relative motion (e.g., position/rotation/direction change, velocity information, and/or acceleration information) based on input 9192. It may include a relative motion module 9120 including -9126). Additionally or alternatively, chip 9100 includes a machine vision module 9130 that includes circuitry 9132, 9136 for analyzing image data provided as input 9192 to detect and/or classify objects. can do. Additionally or alternatively, chip 9100 may be configured to generate an environment digital twin (e.g., a digital representation of the environment), retrieve a stored environment digital twin, and/or update the environment digital twin, and determine the location of the mobile system. Determine the location (e.g., location within the environment or environmental digital twin), determine the pose of the mobile system (e.g., the arrangement of one or more wheels, rims, mechanisms, appendages, or other mobile system components), and output (e.g., 9194) may include an orientation module 9140 that includes circuits 9142 and 9148 for determining orientation information for transmission. The functionality of the various circuits of modules 9110, 9120, 9130, and/or 9140 are described in greater detail below.

Processing core(s) 9106 may be configured to perform any of the functions attributed to chip 9100, with or without assistance from various modules 9110, 9120, 9130, and/or 9140. May include processing core(s). For example, processing core(s) 9106 may utilize and/or call various modules to perform various functions described herein. Processing core(s) 9106 may include general-purpose and/or special-purpose processors. In embodiments, processing core(s) 9106 may use serial, parallel, and/or other processing techniques to achieve the functionality described herein.

Accordingly, processing core(s) 9106 may perform functions in addition to those provided by various modules 9110, 9120, 9130, and/or 9140. For example, processing core(s) may receive the output of one module (e.g., positions output by location module 9110) and input it to another module (e.g., orientation module 9140). It can be provided as. Processing core(s) 9106 may also process the output of any of the module(s) to convert the output to a different format.

Processing core(s) 9106 may also compare data output by different modules for error checking and/or to improve accuracy. For example, if position module 9110 indicates that the position of the system has changed, but relative motion module 9120 indicates that the position of the system has not changed (e.g., because the position signal is or may be inaccurate due to inaccuracies of GPS at a granular level), processing core(s) 9106 may discard and/or modify the output of location module 9110.

In embodiments, processing core(s) 9106 may generate data based on the output of different modules. For example, processing core(s) 9106 may determine a velocity vector data structure based on both the current position output by position module 9110 and the relative motion output by relative motion module 9120. . Other outputs from various modules can be combined in a similar way.

In embodiments, processing core(s) 9106 may further operate to store and/or retrieve data to/from storage 9150. For example, processing core(s) 9106 may store and retrieve images in image library 9152 (e.g., for use by machine vision module 9130, as described in more detail below). may store and retrieve environmental digital twins 9154 (e.g., as created/updated by orientation module 9140, as described in more detail below), (e.g. , to store and retrieve system specification(s) 9156 (to determine information regarding components of a mobile system), and/or to implement various functions described in this application, intelligence module(s) 9158 ) can be saved and searched. In an embodiment, the processing core(s) may use intelligence modules 9158 (which may include one or more of the artificial intelligence modules 8804 of FIG. 104) to provide intelligent services (as described with respect to FIG. 104). Any of the functions of (8800) can be implemented.

Location module 9110 may receive location signals (e.g., GPS signals, cellular signals, WI-FI signals) and determine location (e.g., GPS coordinates or coordinates within some other real or simulated coordinate system/frame). You can. In some embodiments, position signal capture circuit 9112 receives position signal data from input 9192 and performs initial processing (e.g., demodulation, processing on the position signal data in a buffer) to capture the data from the position signal. storage, initial sanity check, etc.). In some cases (e.g., when a location is being determined within the coordinates of an environmental digital twin), location signal capture circuitry 9112 may capture environmental digital twin 9154 from storage and/or from environmental digital twin circuitry 9142. You can search. Position determination circuitry 9114 can then calculate a position based on the captured position data. For example, position determination circuitry 9114 may use trilateration techniques to calculate GPS coordinates and related data (e.g., accuracy/error data) based on GPS signals received from multiple satellites. As another example, location determination circuitry 9114 may use cellular and/or WI-FI data to determine the location of a mobile system. In embodiments, multiple position signals may be used by position determination circuitry 9114 to improve accuracy. Position output circuitry 9116 may then output position data (e.g., one or more data structures representing coordinates and/or related data) (e.g., to processing core(s) 9106) , which in turn may provide the location data to other modules, output the location data as output 9194, or otherwise process the location data to determine orientation information.

Relative motion module 9120 receives as input 9192 an accelerometer, gyroscope, and/or other relative motion signal, and relative motion data (e.g., position and/or data) for one or more real or simulated points/frames. or change in rotation/direction, velocity data, and/or acceleration data). Motion sensor capture circuitry 9122 receives data signals from motion sensors, such as accelerometers, gyroscopes, etc., and performs initial processing (e.g., demodulation, storage in a buffer, initial sanity check) on the data to capture relative motion data. inspection, etc.) can be performed. In some cases (e.g., when relative motion is being determined relative to an environmental digital twin), motion sensor capture circuitry 9122 may capture environmental digital twin 9154 from storage and/or from environmental digital twin circuitry 9142. You can search. Relative motion determination circuitry 9124 then determines relative motion data (e.g., changes in position/rotation/direction, velocity, angular velocity, acceleration, angular acceleration, etc.) for a given point/frame, whether real or simulated. Relative motion data can be processed using integration techniques, dead reckoning techniques, etc. to create one or more data structures representing Relative motion output circuitry 9126 may then output relative motion data (e.g., to processing core(s) 9106), which in turn may provide relative motion data to other modules, or can be output as output 9194, or alternatively, the relative motion data can be processed to determine orientation information.

In embodiments, machine vision module 9130 may receive images, video, or other vision-related signals (e.g., LIDAR data) and process the data to detect and/or classify objects. Image sensor capture circuitry 9132 receives vision-related signals from input 9192 and performs initial processing on the vision-related signals to capture images or other vision data (e.g., demodulation, storage in a buffer, from video). extraction of images, generation of images based on LIDAR data, etc.) can be performed. Object detection circuitry 9134 may then detect one or more objects appearing in the image or other vision data. For example, object detection circuitry 9134 may use image processing techniques, such as line/edge detection and/or other machine learning techniques, to detect the location of an object in image/vision data. In some embodiments, object detection circuitry 9134 may utilize a machine learning model (e.g., stored as intelligence module 9158) for object detection.

Object classification circuitry 9136 may recognize or otherwise classify objects that appear in images or other vision data. In some embodiments (not shown), object detection circuitry 9134 and object classification circuitry 9136 may be the same circuit. For example, the machine vision module 9130 can detect and recognize/classify objects in image/vision data using deep learning technology. In some embodiments, as shown, machine vision module 9130 may use separate circuitry and different technologies (e.g., different machine learned models) to detect and classify objects.

In some embodiments, machine vision module 9130 may utilize image data stored in image library 9152. For example, machine vision module 9130 and/or processing core(s) 9106 may cause object detection circuitry 9134 and/or object classification circuitry 9136 to be based on training data stored in image library 9152. Thus, it can be trained to recognize/classify objects. Examples of image/object classification are described in greater detail throughout this disclosure. In some embodiments, the trained model may be stored as intelligence module 9158. Thus, for example, chip 9100 may be configured to operate in a particular environment by storing images of objects in an image library 9152 for training purposes, and/or by storing customized intelligence modules 9158 trained for that particular environment. Can be configured to recognize objects.

In embodiments, orientation module 9140 may receive various data from input 9192 and/or data from other modules of chip 9100 and use the various data to determine orientation data regarding the mobile system. It can be handled. In some embodiments, environmental digital twin circuitry 9142 may configure and/or update an environmental digital twin based on input 9192 and/or retrieve a stored environmental digital twin 9154. For example, environmental digital twin circuitry 9142 may use LIDAR data, radar data, sonar data, etc. to determine objects, surfaces, or other environmental features near the mobile system. In some cases, environmental digital twin circuitry 9142 may update stored environmental digital twin 9154 based on data detected from input 9192. For example, although the stored environmental digital twin 9154 indicates that a particular object is at a particular location, the environmental digital twin circuitry 9142 determines that the object (e.g., based on objects classified by a machine vision system) Upon detecting that it is actually in the second location, environmental digital twin 9154 may be updated with accurate location information for the object.

Position determination circuitry 9144 may use a variety of techniques to determine position. For example, the location determination circuitry may compare the environmental digital twin generated by the environmental digital twin circuitry 9142 to a prestored environmental digital twin 9154 (e.g., environmental digital twin 9154) to determine the location of the mobile system. If the twin circuit 9142 detects multiple stationary objects near the mobile system, and the same object is located in a particular room in the pre-stored environmental digital twin 9154, then the positioning circuit 9144 determines whether the mobile system is located in the particular room. You can decide where it is located). In some embodiments, positioning circuitry 9144 may include location data obtained from a location module, relative motion data obtained from a relative motion module, and an object obtained from a machine vision module to accurately determine the current location of the mobile system within a particular environment. The detection and classification data, the environmental digital twin generated by environmental digital twin circuitry 9142, and/or any pre-stored environmental digital twin 9154 may be reconciled. Accordingly, position determination circuitry 9144 may utilize any of the data inputs 9192 and/or data generated by other modules of chip 9100 to provide an accurate determination of the location of the mobile system.

In embodiments, attitude determination circuitry 9146 may determine attitude information based on data associated with a wheel, rim, mechanism, attachment, or other component of the mobile system. For example, based on position and/or relative motion data associated with the various components, pose determination circuitry 9146 may determine whether the mobile system is currently sitting, standing, falling, moving forward, or moving backward. You can decide things like that. Attitude determination circuitry 9146 may compare position and/or relative motion data associated with various components to data within one or more system specifications 9156 to determine current posture information. Accordingly, chip 9100 may be configured to operate with a particular mobile system by storing system specifications 9156 for that mobile system in storage 9150.

In embodiments, orientation circuitry 9148 may be used to generate orientation data for transmission as output 9194 or as part of various data generated by other circuits and/or modules and/or received via input interface 9102. Everything can be handled. For example, orientation circuitry 9148 may format data, place it into various data structures, manipulate data, check data for errors, and perform other such functions prior to transmission as output 9194. .

110 illustrates network enhancement chips 9200, one or more of which may be used to enhance the operation and/or performance of communications network(s) as described herein. Chip 9200 may be used by any value chain entity utilizing a communications network. In embodiments, chip(s) 9200 may use artificial intelligence (AI) and other technologies to analyze, predict, optimize, and reconfigure communication network(s). In some of these embodiments, network enhancement chip 9200 utilizes a network digital twin to analyze, predict, optimize, and reconfigure the network (e.g., create, access, update, process, render, and/or otherwise use) can be used. A network digital twin can provide a virtual representation of the physical communication network(s) that a network device can access and the current state of that network(s) and/or the network device, as described in more detail below. For example, a network digital twin is an available communication network for a device or set of devices (e.g., LAN network, WIFI network, cellular network (e.g., 4G, 5G, etc.), satellite network, Bluetooth network, RFID a set of devices (networks, etc.), a device or each network to which each device is or has been connected in the past, real-time data associated with each of each network (e.g., current data flow, current bandwidth metrics, current throughput metrics, current error rate) , current traffic type, etc.), historical data related to each network (e.g., historical data flow, historical bandwidth metric, historical throughput metric, historical error rate, historical traffic type, etc.). In an embodiment, network enhancement chip 9200 may, for example, predict which configuration of the network may optimize certain network characteristics and then reconfigure host devices and/or other devices on the network accordingly (e.g. For example, you can switch protocols, switch networks, configure schedules for transmission of data, configure data priorities, configure compression of specific data, configure reformatting of specific data, and Such information may be used to optimize the network (configure upsampling and/or downsampling, configure dropping, buffering, or scheduling of specific data, etc.).

As described herein, chip(s) 9200 receives one or more inputs 9292 from one or more network(s) and analyzes, predicts, and optimizes the network(s) based on the inputs 9292. , and may perform one or more AI-assisted functions to configure. In embodiments, input 9292 may be configured to include network signals (e.g., traffic data and/or data from other network devices) and/or information about the network signals (e.g., signal strength or other characteristics of the network signals). may include. Chip(s) 9200 may then determine and transmit output 9294 containing instructions to optimize or otherwise reorganize the network and/or data communicated thereon. Chip(s) 9200 may be part of a host device (e.g., server device, client device, router device, etc.), which may be anywhere within the network and/or may be a virtual device hosted on a hardware device. That is, the host device may include any device connected to a communication network.

In embodiments, network enhancement chip 9200 generates network-specific data by analyzing one or more connected communication network(s) and can be used to generate network-specific data from other components of the host device, from other network devices, and/or from other network enhancement chip(s). Configured to receive network specific data from 9200. Network enhancement chips may use (e.g., analyze or otherwise leverage) network-specific data to update information about the communications network (e.g., update the network digital twin) and predict future conditions of the network. there is.

In embodiments, network enhancement chip 9200 may analyze network traffic data at various levels of granularity. For example, a network enhancement chip may analyze traffic flows and/or individual data messages (e.g., packets) based on message headers and/or message payloads. Additionally or alternatively, network enhancement chip 9200 may receive messages from other network enhancement chip(s) 9200 and/or network devices. These messages may provide device information that can be used by network enhancement chip 9200 to create and/or update a network digital twin.

In an embodiment, network enhancement chip 9200 determines network-specific data (e.g., data indicative of the quality/reliability of one or more network links) and determines future network conditions (e.g., data indicative of the quality/reliability of one or more network links) You can analyze the physical properties of network signals, such as signal strength, packet error rate, retransmission, etc. Network enhancement chip 9200 may use this information to create and/or update a network digital twin.

In embodiments, network enhancement chip 9200 may be configured to determine the current state of the network, the past state of the network, or the future predicted state of the network (e.g., historical network data metrics, predicted network demand), as described in more detail below. , one or more AI-enhanced technologies may be used to determine optimization for the network based on the network (represented by the network digital twin, etc.). Accordingly, network enhancement chip 9200 may determine optimizations for traffic flows in the network, specific types or configurations of data carried on the network, messages on the network, and/or devices on the network, and expected effects of such optimizations. .

Network enhancement chip 9200 may then initiate and/or perform network optimization. For example, network enhancement chip 9200 may reconfigure a network or segment (e.g., by performing traffic shaping or otherwise modifying data flows or other data received as input 9292) and/or other devices. may be configured to instruct the network or segment to reconfigure the network or segment.

Network enhancement chip 9200 may initiate reconfiguration of the network, traffic flow on the network, data transmitted over the network, devices on the network, etc., as described in more detail below. In embodiments, network enhancement chip 9200 may instruct one or more network devices to perform one or more reconfiguration functions to cause optimization in the network. Additionally or alternatively, network enhancement chip 9200 may be used to reroute flows (e.g., switching from one network to another and/or switching routing paths on a network), formatting flows, and/or reconfigure the network by changing protocols, or otherwise modifying flows.

In embodiments, network enhancement chip 9200 may reconstruct data transmitted over a network by processing the data according to one or more optimizations. For example, network enhancement chip 9200 is configured to compress or decompress data, reformat data, resample data, bundle data, schedule data transmission of bundled data, etc. It can be.

In embodiments, chip(s) 9200 may be modular component(s) that can be integrated with one or more networks (e.g., as a stand-alone device) and/or network device(s) in a variety of ways. For example, multiple network devices may each include a network enhancement chip 9200 that may communicate with each other to exchange information, determine optimization, and/or configure the network at various points in the network. To facilitate modularity, chip(s) 9200 may be provided partially or completely within a housing (not shown) and may be provided with electrical connectors, optical connectors, and/or wireless connectors (e.g., antenna, inductive coil, etc.) may receive an input 9292 and/or provide an output 9294. Additionally or alternatively, chip(s) 9200 may be integrated with other circuits, processors, systems, etc. on one or multiple substrates/chips.

Chip(s) 9200 may include one or more system-on-chip (SOC), integrated circuit (IC), application specific integrated circuit (ASIC) components to provide the functionality and/or any other functionality ascribed to chip 9200. It may be and/or may include the like. For example, chip 9200 may be provided as part of a SOC that also provides other functionality described in this application. Generally, the components of chip 9200 may include one or more general-purpose processing chips configured using software instructions or other code and/or special-purpose processing chips (e.g., For example, ASIC) may be included.

Multiple chip(s) 9200 may be used to perform the functions described in this application. For example, multiple chip(s) 9200 may be used to perform analysis, optimization, and/or configuration functions more quickly, or to transfer more complex calculations from one chip 9200 to another chip with a better power source. Serial, parallel, and/or other processing techniques may be used, such as offloading to 9200 to perform such functions more efficiently. As another example, one chip 9200 may be used to provide network enhancement functionality for one portion of a network (e.g., a specific area covered by a wireless network) while another chip 9200 may be used to provide network enhancement functionality for a portion of the network (e.g., a specific area covered by a wireless network). may be used to provide network enhancement functionality for a second portion of a network (e.g., a different area covered by the same wireless network).

In an embodiment, physical input interface 9202 receives one or more inputs 9292 to network enhancement chip 9200 as described herein. Input 9292 may be transmitted to physical input interface 9202 via one or more physical network(s) by another network device that may or may not include corresponding network enhancement chip(s) 9200. Physical network(s) may include any type of wired or wireless network. Input 9292 may provide information about network traffic, information about the network, information about network devices, and information about optimizing or otherwise configuring the network (e.g., as received from other network enhancement chip(s) 9200). It may contain one or more of commands, etc. Input 9292 may also include data that may be stored in storage 9250, such as a protocol for protocol library 9252, a network digital twin 9254 (e.g., a digital representation of a network), one or more system specifications ( s) 9256, and/or one or more intelligence module(s) 9258.

As noted above, output data 9294 transmitted from physical output interface 9204 may be used to store network traffic, host devices, including network enhancement chips (e.g., for use by other network enhancement chips 9200). and/or instructions for optimizing or otherwise configuring the network (e.g., to be transmitted to other network devices and/or network enhancement chip(s) 9200). In embodiments, the output of chip 9200 may be transmitted by physical output interface 9204 over any of the physical network(s) connected to the host device.

In an embodiment, chip 9200 may include one or more of network analysis module 9210, optimization module 9220, data configuration module 9230, and network configuration module 9240. In an embodiment, network analysis module 9210 may include circuitry 9212-9216 to analyze a network and/or create/update a network digital twin based on input 9292. Additionally or alternatively, chip 9200 includes an optimization module 9220 that includes circuitry 9222, 9228 for predicting one or more optimizations for a network based on input 9292 and/or a network digital twin. can do. Additionally or alternatively, chip 9200 may include circuitry 9232, 9236 for configuring/optimizing network data received as input 9292 and transmitting the configured/optimized network data as output 9294. It may include a configuration module 9230. Additionally or alternatively, chip 9200 may receive a traffic flow as input 9292, configure/optimize the traffic flow, send instructions to another network device to effect configuration/optimization of the traffic flow, and configure/optimize the traffic flow. /may include a network configuration module 9240 that includes circuitry 9242, 9246 for outputting optimized traffic flows and/or instructions as output 9294. The functionality of the various circuits of modules 9210, 9220, 9230, and/or 9240 are described in more detail below.

Processing core(s) 9206 may be configured to perform any of the functions attributed to chip 9200, with or without assistance from various modules 9210, 9220, 9230, and/or 9240. May include processing core(s). For example, processing core(s) 9206 may utilize and/or call various modules to perform various functions described herein. Processing core(s) 9206 may include general-purpose and/or special-purpose processors. In embodiments, processing core(s) 9206 may use serial, parallel, and/or other processing techniques to achieve the functionality described herein.

Accordingly, processing core(s) 9206 may perform functions in addition to those provided by various modules 9210, 9220, 9230, and/or 9240. For example, processing core(s) may receive the output of one module (e.g., optimization determined by optimization module 9220) and transfer it to another module (e.g., data organization module 9230 and/or It can be provided as input to the network configuration module 9240). Processing core(s) 9206 may also process the output of any of the module(s) to convert the output to a different format.

In embodiments, processing core(s) 9206 may further operate to store and/or retrieve data to/from storage 9250. For example, processing core(s) 9206 may store and retrieve protocols in protocol library 9252 (e.g., for use by various modules, as described in more detail below); store and retrieve network digital twins 9254 (e.g., as created/updated or otherwise utilized by various modules, as described in more detail below); store and retrieve system specification(s) 9256 (to determine information about network devices), and/or store intelligence module(s) 9258 to implement various functions described herein. and searchable. In an embodiment, the processing core(s) may use intelligence modules 9258 (which may include one or more of the artificial intelligence modules 8804 of FIG. 104) to provide intelligent services (as described with respect to FIG. 104). Any of the functions of (8800) can be implemented.

Network analysis module 9210 may detect network signals (e.g., network traffic between various network endpoint devices, messages containing information about network devices, etc.), information about network signals (e.g., signals of network signals, etc. strength or other physical properties), and/or other network information (e.g., data indicative of current or historical network performance, current or historical network device information, network digital twin(s) generated by other devices, etc.) In addition to receiving and determining information about a network, it may also create and/or update one or more network digital twin(s) corresponding to various communication network(s).

In embodiments, signal analysis circuitry 9212 may receive a network signal from input 9292 and perform signal analysis (e.g., analysis of header information and/or payload information) to determine information about the signal. . For example, signal analysis circuitry 9212 may analyze header information (e.g., to/from an address, protocol, flow identifier, etc.) and/or payload information (e.g., type of data included in the payload, data Based on whether the network traffic is encrypted, etc., it can be analyzed whether the network traffic belongs to a specific traffic flow. As another example, signal analysis circuitry 9212 may detect a message containing device information regarding a network device. Additionally or alternatively, signal analysis circuitry 9212 may analyze physical properties of the signal received as input 9292, such as signal strength indicators. In such embodiments, signal analysis circuitry 9212 may analyze additional physical properties over time (e.g., to determine that signal strength is weakening and/or predict that a corresponding wireless link is likely to be lost). It can be analyzed. Signal analysis circuitry 9212 may analyze all or only a portion of any network traffic received as input 9292. For example, a signal analysis circuit may sample one of every N network packets received as input 9292, analyze the physical properties of the signal every N microseconds, etc.

In embodiments, data analysis circuitry 9214 may determine additional network information based on data in the analyzed signal. For example, data analysis circuitry 9214 may determine whether the data for a particular traffic flow is encrypted, compressed, has a particular format, and is associated with a particular priority level (e.g., a priority level associated with the contracted data rate). You can analyze whether it works, etc. Data analysis circuitry 9214 may add such information to one or more corresponding network digital twins, each of which may represent a specific communications network carrying the data, one or more devices on the network, one or more data configurations for the network, and one or more data configurations for the network. You can specify one or more rate schedules, etc. In embodiments, data analysis circuitry 9214 may be application-specific, which may indicate a particular application and/or one or more attributes (e.g., whether the data is payment data, customer data, whether the data is associated with a particular project, etc.). Data can be analyzed. In this embodiment, data analysis circuitry 9214 may add this information to the network digital twin. Additionally or alternatively, data analysis circuitry 9214 may analyze the received message to detect information regarding the network device. For example, data analysis circuitry 9214 may analyze data included in the data message (e.g., MAC address or other identifier) to identify the specific manufacturer, model, or identity of the network device. In these cases, data analysis circuitry 9214 may then retrieve additional information regarding the identified network device using system specifications 9256 corresponding to the identified device. Additionally or alternatively, data analysis circuitry 9214 may analyze status messages indicating the current state of the network device, such as battery level, currently available bandwidth, currently available processing capacity, etc. The data analysis circuit 9214 may store information about various network devices in the network digital twin 9254 corresponding to a specific network.

In embodiments, network diagnostic circuitry 9216 may then determine network information based on the analyzed signals and/or data. For example, network diagnostic circuitry 9216 may determine the protocol, format, endpoint device, bandwidth, and/or throughput (e.g., current, average, minimum, and/or maximum) for each traffic flow on one or more connected networks. bandwidth/throughput), error rate, packet loss rate, flow priority, flow quality of service (QoS) metrics/requirements, flow schedule, application-specific data, etc. can be detected and recorded. As another example, network diagnostic circuitry 9216 may detect a new traffic flow and add it to a list of traffic flows for a particular network. Network diagnostic circuitry 9216 may also determine diagnostic information indicative of errors or other conditions in the network. For example, if network diagnostic circuitry 9216 detects that no traffic is being received over a particular network or from a particular device, it may detect that the corresponding network/device is not available. In embodiments, network diagnostic circuitry 9216 may perform diagnostic workflows to detect problems or other conditions on the network. For example, network diagnostic circuitry 9216 may poll network devices for status information, attempt to transmit data over one or more communication networks, send or receive test data flows to measure bandwidth, throughput, etc. , and may perform other such diagnostic functions. In embodiments, network diagnostic circuitry 9216 may use the determined network/diagnostic information to create or update one or more network digital twins 9254 corresponding to specific networks, network devices, data configurations, rate schedules, etc.

In embodiments, optimization module 9220 utilizes network analysis output by network analysis module 9210 and/or network digital twin(s) 9254 to determine one or more optimizations and their predicted effects on the network. You can decide. Optimization module 9220 may use AI-assisted functions (e.g., , machine learning models or other intelligence modules 9258) may be used.

In embodiments, data optimization circuitry 9222 may predict the effect of one or more optimizations to be applied to network data. For example, data optimization circuitry 9222 may utilize intelligence module 9258 (e.g., a trained deep learning model) and/or stored optimization parameters to optimize specific network metrics based on current network information. It may be determined that certain types of data should be rerouted (e.g., over a different network), compressed, downsampled, dropped, buffered, and/or rescheduled. Optimization parameters may be specified by one or more system specifications 9256, such that data optimization circuitry 9222 may be configured to optimize the communication network in a specified manner by storing the corresponding system specifications in storage 9250. As another example, data optimization circuitry 9222 may cause the network to improve the quality of data transmitted over the network, such as by upsampling, decompressing, providing a higher priority, or otherwise increasing the quality of data transmitted over the network. AI-assisted technology (e.g., utilizing intelligence module 9258) may be used to determine that there is sufficient capacity to increase quality. In this example, system specifications 9256 state that data optimization circuitry 9222 should optimize flows, data types, transmit/receive network devices, etc. for increased data quality generally, increased data quality for specific applications, etc. can indicate that Accordingly, data optimization circuitry 9222 may utilize AI techniques to optimize various network parameters as required by specific system specifications.

In embodiments, network optimization circuitry 9224 may determine one or more optimizations to apply to a network device. For example, network optimization circuitry 9224 may utilize intelligence module 9258 (e.g., a trained deep learning model) to determine, based on current network information, certain network devices to take certain actions (e.g., power Up or down, switching networks, adjusting the transmission schedule of other devices, adjusting the protocols used by other network devices, rerouting traffic from other devices, transmitting or receiving by other devices. that network performance should be improved (e.g., by performing compression or some other data modification on all traffic being generated) or optimized for some other parameter (e.g., as indicated by the system specification 9256). You can decide. Similarly, network optimization circuitry 9224 may determine that a network device should take certain actions to improve the quality of data transmitted over the network and/or perform any other optimizations.

In embodiments, data optimization circuitry 9222 and/or network optimization circuitry 9224 may utilize network security circuitry 9226 and/or network governance circuitry 9228 as part of determining optimization for data and/or network. You can utilize it. Network security circuitry 9226 may enforce security rule(s) that may change and/or override optimizations proposed by data optimization circuitry 9222 and/or network optimization circuitry 9224. For example, network security circuitry 9226 may be used to control network device(s), data, and/or network(s) to determine whether a proposed optimization is sufficiently or insufficiently secure, or otherwise complies with security rules. ) can be analyzed for the proposed optimization. As a specific example, network security circuitry 9226 may determine that a proposed optimization involving decryption of network traffic may be insecure for certain data types or traffic flows, and may therefore override and/or change the proposed optimization. there is.

In a similar manner, network governance circuit 9228 may enforce governance rules specifying specific legal requirements, business requirements, technical requirements, etc. Accordingly, network governance circuitry 9228 may change and/or override optimizations proposed by data optimization circuitry 9222 and/or network optimization circuitry 9224 such that the optimization complies with the governance rule(s). In embodiments, network security circuitry 9226 and/or network governance circuitry 9228 may utilize intelligence module 9258 to store and/or otherwise specify security and/or governance rules. In embodiments, network security circuitry 9226 and/or network governance circuitry 9228 may use intelligence module 9258 (which may include one or more of artificial intelligence modules 8804 of Figure 104) to Any of the functionality of intelligence service 8800 (as described in connection with) may be implemented.

In an embodiment, data composition module 9230 receives data traffic via input 9292 and optimizes data traffic by data optimization circuitry 9222 and/or network optimization circuitry before sending it as output 9294. Any optimization determined by 9224 may be applied to the received data traffic. Data capture/extraction circuitry may receive inbound or outbound data packets (e.g., from another network device and/or from a host device) and may extract data from the data packets.

The data encryption circuit 9234 may perform any necessary encryption/decryption operations on the extracted data. Data encryption circuitry 9234 may decrypt data received from another device so that the data can be analyzed and data-specific optimizations applied. For example, if the optimization module 9220 indicates that the data should be reformatted (e.g., up or downsampled, compressed/decompressed, etc.), the data must first be processed through the data encryption circuit 9234 before optimization can be applied. ) may need to be decrypted. Additionally or alternatively, data encryption circuitry 9234 may encrypt data if optimization module 9220 determines that data encryption should be applied (e.g., to increase the security of a particular type of data or traffic flow). can be applied.

Data processing circuitry 9236 may perform arbitrary processing on data to implement optimizations determined by optimization module 9220. For example, data optimization circuitry 9222 and/or network optimization circuitry 9224 may be configured to determine certain attribute(s) (e.g., a specific type of data, a specific data flow, a specific application-specific attribute, a specific data priority, When you determine that data associated with a particular data protocol (e.g., a specific data protocol, etc.) should be optimized by being processed in a particular way (e.g., by compression/decompression, upsampling or downsampling, reformatting, delay, buffering, rescheduling, etc.), , the data processing circuit may perform the corresponding processing when detecting data matching the attribute(s). Accordingly, data processing circuitry 9236 may perform data optimization on data received by network enhancement chip 9200.

In embodiments, network configuration module 9240 may transmit and receive signals to and from a communications network to perform certain optimizations for the network and/or network devices as determined by network optimization circuitry 9224. The network configuration module 9240 may perform network optimization in parallel or sequentially before or after the data configuration module 9230 optimizes data.

In an embodiment, signal processing circuitry 9242 may be configured to coordinate with other network enhancement chip(s) 9200 and/or network devices on a network (e.g., to/from and/or network enhancement chip(s) 9200 on a network. 9200) may generate and receive inbound or outbound data signals) to/from a host device. For example, network enhancement chip 9200 may send a signal to a target network device instructing the target network device to perform some action (e.g., as determined by network optimization circuitry 9224) to optimize the network. Can be sent. Additionally or alternatively, signal processing circuitry 9242 may receive instructions from other network enhancement chip(s) on the network directing network enhancement chip 9200 to perform configuration functions to optimize the network.

In embodiments, signal processing circuitry 9242 may modify signals transmitted to other network devices based on optimization determined by optimization module 9220. For example, signal processing circuitry 9242 detects a message being sent (e.g., by another network device) that will cause the target network device to use a first protocol, but optimization module 9220 detects a message being sent (e.g., by another network device) that will cause the target network device to use the first protocol. If it is determined that using a protocol will optimize the network, then signal processing circuitry 9242 may modify the message to indicate use of a second protocol instead. Similarly, signal processing circuitry 9242 may drop (e.g., delete without sending) or delay a message being transmitted to another device if the message contains instructions that conflict with optimizations determined by optimization module 9220. there is. Accordingly, signal processing circuitry 9242 may effect optimization by delaying or overriding various instructions transmitted and received by various network devices.

In embodiments, protocol conversion circuitry 9244 may configure the protocol of data signals transmitted over a network. As a specific example, protocol conversion circuitry 9244 may convert specific types of data or data flows from the TCP/IP protocol to the UDP/IP protocol to optimize specific network parameters. Protocol conversion circuitry 9244 may access protocol information from protocol library 9252 to configure one or more protocols. In embodiments, protocol conversion circuitry 9244 may reconfigure other protocol level attributes of signals and/or other data to be transmitted over the network. For example, protocol conversion circuitry 9244 may reconstruct the source or destination address, protocol time stamp, protocol stream identifier, and/or any other fields of the protocol header. Additionally or alternatively, protocol conversion circuitry 9244 may generate instructions for transmission to another network device that may cause the other network device to reconfigure the protocol of a data signal transmitted by that network device. Protocol switching circuitry 9244 may reconfigure the protocol of traffic on the network based on optimizations determined by optimization module 9220. Additionally or alternatively, the protocol switching circuitry 9244 may perform a switch based on the current state of the network (e.g., as indicated by the network digital twin 9254) and/or by the signal processing circuitry 9242. The protocol can be reconfigured based on processing.

In embodiments, network switching circuitry 9246 may reconfigure routing, scheduling, network topology, or other properties of one or more network(s) communicating with network enhancement chip 9200. For example, network transition circuitry 9246 may reconfigure a network from a mesh topology to a star topology (e.g., by instructing one or more network devices to change roles) and reconfigure the network to a star topology (e.g., by instructing one or more network devices to change roles) and routing traffic across one network instead of another (to balance bandwidth, for example); routing traffic through one router instead of another (for example, to balance the load on two routers); Scheduling the transmission of the first traffic in 1 transmission slot and the second traffic in the second transmission slot can be performed, etc. In some cases, network switching circuitry 9246 may reconfigure the routing and/or scheduling of data received by network enhancement chip 9200 (e.g., as input 9292). Additionally or alternatively, protocol switching circuitry 9244 may generate instructions for transmission to another network device that may cause the other network device to reconfigure aspects of the network. Network transition circuitry 9246 may reconfigure the network based on optimizations determined by optimization module 9220. Additionally or alternatively, the network transition circuitry 9246 may be configured to provide a signal processing circuit 9242 based on the current state of the network (e.g., as indicated by the network digital twin 9254) and/or The network can be reconstructed based on prediction/analysis.

111 illustrates diagnostic chips 9300, one or more of which may be used to perform one or more diagnostic functions as described herein. Chip 9300 may be used by any value chain entity to perform diagnostics. In embodiments, chip(s) 9300 may perform diagnostics based on data from one or more sensors, including biological sensors, chemical sensors, and/or electromechanical sensors, and analyze and recommend actions based on the diagnostics. Artificial intelligence (AI) and other technologies may be used to generate reports, including: In embodiments, diagnostic chip 9300 performs one or more specific diagnostics by receiving, storing, and utilizing corresponding specifications indicating the type of sensor input, how to process and format sensor input, how to analyze sensor input, etc. It can be configured to do so. Similarly, diagnostic chip 9300 may be configured to perform a particular diagnosis by receiving, storing, and utilizing corresponding analysis libraries and/or intelligence modules that may be used to configure and perform one or more analyses.

In embodiments, chip 9300 may include, without limitation, chemical sensors (e.g., hazard-specific sensors, flammability sensors, compound-specific sensors, etc.), biological sensors (e.g., bio-hazardous materials and/or hazardous level sensors, radiation sensors, etc.), electromechanical sensors (e.g. vibration sensors, stress/strain sensors, electrical resistance/current sensors, sensors that measure motion and/or position data such as inertia, velocity, acceleration, GPS, etc.) , a wide variety of sensors, including optical/imaging (e.g., optical sensors, hyperspectral sensors, intensity sensors, thermal sensors, etc.) and other environmental sensors (e.g., temperature sensors, humidity sensors, air movement sensors, etc.). Can be configured or reconfigured to receive and interpret data from. Chip 9300 may be reconfigured to receive and interpret specific sensor data based on sensor specification(s) that enable chip 9300 to receive and interpret sensor data from a corresponding sensor.

In embodiments, chip 9300 may be configured or reconfigured to perform organic analysis, laboratory analysis, and/or electromechanical analysis based on sensor data. For example, chip 9300 may be a lab-on-chip and/or organ-on-chip that may allow for simulating organisms, performing laboratory analyses, performing electromechanical analyses, etc. -chip) functionality may be included. Chip 9300 performs the corresponding simulation/analysis and makes predictions using corresponding AI technology (e.g., using a deep learning model trained to interpret corresponding sensor data). Receive, store, and utilize certain analysis libraries and/or intelligence modules that allow you to do things like: Using similar techniques, chip 9300 may further combine the results of various analyzes to perform one or more combined analyzes.

In embodiments, chip 9300 may be configured to use a governance library to control analytics, perform predictions, and/or provide recommendations. For example, a governance library can indicate whether a particular condition is acceptable and thus control whether action should be taken to resolve the condition. Chip 9300 may be configured to report the results of any analysis, including current or predicted conditions, recommended actions to address the conditions, etc.

In embodiments, chip(s) 9300 may be modular component(s) that can be integrated with a host system in a variety of ways. For example, the chip(s) may be integrated with a mobile host system (eg, a robot), a stationary host system, or any other host system that receives sensor input. To facilitate such modularity, chip(s) 9300 may be provided partially or completely within a housing (not shown) and may include electrical connectors, optical connectors, and/or wireless connectors (e.g., antennas, Induction coil, etc.) may receive an input 9392 and/or provide an output 9394. Additionally or alternatively, chip(s) 9300 may be integrated with other circuits, processors, systems, etc. on one or multiple substrates/chips.

Chip(s) 9300 may include one or more system-on-chip (SOC), integrated circuit (IC), application specific integrated circuit (ASIC) components to provide functionality and/or any other functionality ascribed to chip 9300. It may be and/or may include the like. For example, chip 9300 may be provided as part of a SOC that also provides other functionality described in this application. Generally, the components of chip 9300 may include one or more general purpose processing chips configured using software instructions or other code and/or special purpose processing chips (e.g. For example, ASIC) may be included.

Multiple chip(s) 9300 may be used to perform the functions described in this application. For example, multiple chip(s) 9300 may be used to perform analysis more quickly, by offloading more complex calculations from one chip 9300 to another chip 9300 with a better power source. Serial, parallel, and/or other processing techniques may be used to perform processing more efficiently. As another example, one chip 9300 may be used to provide a first analysis and a second analysis, while the other chip 9300 may be used to provide a combined analysis based on the first analysis and the second analysis. can be used

In an embodiment, physical input interface 9302 receives one or more inputs 9392 to diagnostic chip 9300 as described herein. Input 9392 may be connected to physical input interface 9302 by other chips, circuits, modules, and/or other components of the host system, or by other devices in communication with the host system (e.g., over a communications network). can be sent. For example, input data may come from a sensor, sensor processing chip/module/circuit, antenna, storage device, network interface, or any other data source for chip(s) 9300 as described herein. You can. Physical input interface 9302 may be coupled to the source(s) of input 9392 via a wired or wireless connection. As in the states above, input 9392 may include any type of sensor data. Input 9392 may also include data that may be stored in storage 9350, such as analysis rules/configuration for analysis library 9352, governance rules/configuration for governance library 9354, one or more system specification(s). 9356 (e.g., sensor specifications), and/or one or more intelligence module(s) 9358.

Output data 9394 transmitted from physical output interface 9304 may include results of the analysis, specific conditions indicated by the analysis, predictions, other diagnostic information, and/or recommended actions to address any specific or predicted conditions. May include report(s) that indicate In embodiments, the output of chip 9300 may be transmitted by physical output interface 9304 to other chips, circuits, modules, and/or other components of a host system or other devices in communication with the host system, as described herein. can be sent. Physical output interface 9304 can connect to these components via wired or wireless connections.

In embodiments, chip 9300 may include one or more of sensor module 9310, analysis module 9320, and/or output module 9330. In an embodiment, sensor module 9310 may include circuitry 9312-9318 to receive sensor data received as input 9392 and perform initial processing (e.g., filtering). Additionally or alternatively, chip 9300 may include circuitry 9322 to perform analysis, detect conditions, predict future conditions, generate other diagnostic information, and generate recommendations for addressing any conditions. It may include an analysis module 9320 including 9326). Additionally or alternatively, chip 9300 may include circuitry 9332, 9336 to perform additional combined analysis, enforce governance rules for analysis, predictions, recommendations, etc., and output reports containing diagnostic/analytical data. It may include an output module 9330 including. The functionality of the various circuits of modules 9310, 9320, and/or 9330 are described in greater detail below.

Processing core(s) 9306 may be one or more processing cores that may be configured to perform any of the functions attributed to chip 9300, with or without assistance from various modules 9310, 9320, and/or 9330. May include (s). For example, processing core(s) 9306 may utilize and/or call various modules to perform various functions described herein. Processing core(s) 9306 may include general-purpose and/or special-purpose processors. In embodiments, processing core(s) 9306 may use serial, parallel, and/or other processing techniques to achieve the functionality described herein.

Accordingly, processing core(s) 9306 may perform functions in addition to those provided by various modules 9310, 9320, and/or 9330. For example, processing core(s) may receive the output of one module (e.g., sensor data output by sensor module 9310) and transmit it to another module (e.g., analysis module 9320). It can be provided as input. Processing core(s) 9306 may also process the output of any of the module(s) to convert the output to a different format.

In embodiments, processing core(s) 9306 may further operate to store and/or retrieve data to/from storage 9350. For example, processing core(s) 9306 may store analysis configurations/data and/or data in analysis library 9352 (e.g., for use by analysis module 9320, as described in more detail below). Alternatively, governance configurations/data may be stored and retrieved in governance library 9354 and system specifications 9356 (e.g., to configure sensor module 9310, as described in more detail below). and retrieve, and/or store and retrieve intelligence module(s) 9358 to implement various functions described herein. In an embodiment, the processing core(s) may use intelligence modules 9358 (which may include one or more of the artificial intelligence modules 8804 of FIG. 104) to provide intelligent services (as described with respect to FIG. 104). Any of the functions of (8800) can be implemented.

Sensor module 9310 may receive and perform initial processing on sensor data from any type of sensor. In some embodiments, biological sensing circuitry 9312 may receive and/or process (e.g., filter, sanity check, error check, etc.) sensor data from biological sensors. Additionally or alternatively, chemical sensing circuitry 9314 may receive and/or process (e.g., filter, sanity check, error check, etc.) sensor data from chemical sensors. Additionally or alternatively, electromechanical sensing circuitry 9316 may receive and/or process (e.g., filter, sanity check, error check, etc.) sensor data from electrical sensors, mechanical sensors, and/or electromechanical sensors. can do. Additionally or alternatively, environmental sensing circuitry 9318 may receive and/or process (e.g., filter, sanity check, error check, etc.) sensor data from environmental sensors, including atmospheric sensors, imaging sensors, etc. there is.

In embodiments, biological sensing circuit 9312, chemical sensing circuit 9314, and/or electromechanical sensing circuit 9316 each have system specifications corresponding to a particular sensor to configure the sensing circuit to process the corresponding sensor data. (9356) can be accessed. For example, when diagnostic chip 9300 is configured to perform a particular organic analysis (e.g., prediction/simulation/testing of a particular organ or organ system), biological sensing circuit 9312 may be configured to perform a specific organic analysis (e.g., prediction/simulation/testing of a particular organ or organ system). For example, a system specification 9356 can be retrieved for a microfluidic sensor, bioMEMS sensor, etc.) so that the biological sensing circuit 9312 can receive and process relevant sensor data (e.g., format, filter, error check, etc.) can do. As another example, when diagnostic chip 9300 is configured to perform a particular laboratory analysis (e.g., drug test, disease test, etc.), chemical sensing circuit 9314 may detect a corresponding sensor (e.g., chemical sensor). The system specifications 9356 can be retrieved to allow the chemical sensing circuit 9314 to receive and process relevant sensor data (e.g., format, filter, error check, etc.). As another example, when the diagnostic chip 9300 is configured to perform electromechanical analysis (e.g., diagnostic analysis of a particular machine/circuit based on vibration sensors, electrical sensors, electromechanical sensors, etc.), the electromechanical detection circuit 9316 ) can retrieve system specifications 9356 for a corresponding sensor (e.g., MEMS sensor, vibration sensor, etc.) so that electromechanical sensing circuitry 9316 can receive and process relevant sensor data (e.g., format, filter, error checking, etc.). As another example, when diagnostic chip 9300 is configured to perform environmental analysis (e.g., diagnostic analysis based on imaging data and/or environmental data), environmental sensing circuitry 9318 may detect a corresponding sensor (e.g. , imaging sensors, optical sensors, other environmental sensors, etc.) so that environmental sensing circuitry 9318 can receive and process relevant sensor data (e.g., format, filter, error check, etc.) etc.) can be done.

Analysis module 9320 receives processed sensor data from sensor module 9310 and performs various analyzes using organic analysis circuitry 9322, laboratory analysis circuitry 9324, and/or electromechanical analysis circuitry 9326. can do. Organic analysis circuitry 9322, laboratory analysis circuitry 9324, electromechanical analysis circuitry 9326, and/or environmental analysis circuitry 9328 each include an analysis library 9352 and/or an intelligence module ( You can retrieve the analysis configuration(s) from s) (9358). For example, when diagnostic chip 9300 is configured to perform a simulation of a particular organ or organ system, organic analysis circuitry 9322 may determine configuration parameters corresponding to the organ/organ system (e.g., specific biology, functional mechanism, etc.). 9352), and trained to predict and/or analyze the response of an organ/organ system to physiological stimuli, specific drugs, specific diseases, and/or other inputs. You can search for the intelligent module (9358). Similarly, when the diagnostic chip 9300 is configured to perform a disease test/analysis, the laboratory analysis circuitry 9324 may be configured with an analysis library (e.g., specific indicators, symptoms, etc.) that specifies configuration parameters corresponding to the disease. Analysis data can be retrieved from 9352) and an intelligence module 9358 trained to predict disease progression, disease response to treatment, etc. Similarly, when diagnostic chip 9300 is configured to perform diagnostic analysis of a machine, electromechanical analysis circuitry 9326 may display configuration parameters for the machine (e.g., a particular state of the machine or sub-portion of the machine). Analytical data may be retrieved from the analysis library 9352 (e.g., frequencies and/or frequency patterns, electrical information indicative of accurate or incorrect operating levels for the machine's electrical circuits, etc.) and may cause a potential failure of the machine or other Intelligence modules 9358 trained to predict conditions, effects of maintenance actions, etc. can be retrieved. Similarly, when the diagnostic chip 9300 is configured to perform a diagnostic environmental analysis, the environmental analysis circuitry 9328 may determine configuration parameters for the environment (e.g., image/optical data indicative of specific conditions of the environment and/or Analysis data specifying other environmental data, etc.) can be retrieved from the analysis library 9352, and an intelligence module trained to predict potential environmental conditions, such as safe/unsafe conditions for humans and/or other environmental conditions ( 9358) can be searched.

Organic analysis circuit 9322, laboratory analysis circuit 9324, electromechanical analysis circuit 9326, and/or environmental analysis circuit 9328 each perform analysis, determine/predict conditions, and process/maintenance/ One or more AI assistive technologies may be used to, for example, predict the effectiveness of preventive actions. For example, one of the circuits may configure a first AI assistance technique (e.g., using configuration parameters specified by an analysis library) to detect a specific condition (e.g., a gradient boosted tree model). The same or different circuits can then use different AI assistance technologies (e.g., neural networks trained using deep learning techniques) to predict response to a treatment plan for a specific condition. Similarly, chip 9300 may use multiple AI-assisted technologies to perform the same task to improve the accuracy of diagnostic information. Accordingly, by leveraging multiple AI-assisted technologies, chip 9300 can perform complex and highly accurate workflows utilizing different AI-assisted technologies.

In embodiments, multiple intelligence module(s) 9358 may be used to provide different types of diagnostics for a single workflow. In an embodiment, intelligence module 9358 may include one or more of artificial intelligence modules 8804 of FIG. 104 . Additionally or alternatively, many of the analysis circuits 9322-9328 may be used for analysis workflows. For example, assays for disease diagnostic applications may use both chemical and biological sensors as inputs, and chip 9300 may include organic analysis circuitry 9322 and/or laboratory analysis circuitry to perform aspects of the relevant analysis. 9324) Both can be used correspondingly.

In embodiments, output module 9330 may use results from analysis module 9320 to perform combined analysis, enforce governance rules, and/or use results from analysis module 9320 and/or combined A report containing the results of the analysis generated by the analysis circuitry 9332 may be generated/sent.

Combined analysis circuitry 9332 may correlate and further analyze multiple analyzes generated by analysis module 9320. For example, a first diagnostic analysis (e.g., using a first AI-assisted technology and/or a first set of sensor inputs) indicates the presence of a specific condition (e.g., a disease is present), If a second diagnostic analysis (e.g., using a second AI-assisted technology and/or a second set of sensor inputs) indicates the absence of a particular condition (e.g., a disease is absent), the combined analysis Circuitry 9332 may combine the results of the first and second diagnostic analyzes, apply weights, utilize intelligence module 9358, and/or otherwise process the outputs of the first and second diagnostic analyzes to determine specific You can generate indications such as whether a condition exists or the likelihood that a specific condition exists. In an embodiment, combined analysis circuitry 9332 may process a first diagnostic analysis indicating a first condition and a second diagnostic analysis indicating a second condition to determine that a third condition exists. In an embodiment, combined analysis circuitry 9332 may be configured to determine a first action plan (e.g., treatment/maintenance/prevention action plan) indicated by the first diagnostic analysis and a second action indicated by the second diagnostic analysis. Plans may be combined to include actions indicated by the first action plan, actions indicated by the second action plan, and/or third actions not indicated by the first or second action plan. An action plan can be created. In an embodiment, combined analysis circuitry 9332 processes a first diagnostic analysis indicating a first probability of a condition and a second diagnostic analysis indicating a second probability of a condition to form a combined analysis indicating a third probability of a condition. An analysis may be produced, where the third probability may be lower, higher, in between, or equal to one or both of the first and second probabilities.

Governance circuit 9334 may enforce rules, override actions in an action plan, control the analysis performed by analysis circuits 9322-9328, or direct the analysis and/or output of the analysis to comply with governance rules. You can modify it in other ways. For example, a governance circuit may require that certain actions in an action plan be non-dangerous to humans, non-illegal, etc. Governance circuitry 9334 may retrieve governance rules from governance library 9354, which may store customized rules for specific applications. For example, when chip 9300 is monitoring environmental conditions at a location where a human is working, governance circuitry 9334 may search a governance library that specifies acceptable environmental conditions for humans. Governance circuit 9334 then requests specific actions when certain conditions are detected (e.g., sounding an alarm when a hazardous substance is detected) and, in the action plan, requires specific actions (e.g., placing the environment at risk or This information may be used to override actions that could otherwise be modified to be unsuitable for humans, control which type of analysis is used and/or how the analysis is performed by the various analysis circuits, etc. In contrast, when chip 9300 is configured to monitor environmental conditions in a non-human location, it may use a different set of governance rules. In some cases, governance rules may require reporting certain conditions to certain parties (e.g., reporting disease data to patients, doctors, etc.); You may request a ban on reporting. To control the operation of the analysis circuit, the governance circuit 9334 may instruct the analysis circuit to perform or not perform a particular analysis, modify how the analysis is performed, etc. , may be configured to monitor and/or be utilized by analysis circuits 9322-9328.

Reporting circuitry 9336 may generate a report containing the results of the analysis and/or combined analysis, as modified by any governance rules, and output the report (e.g., as output 9394). You can. Reporting circuitry 9336 may format data as required for interoperability with any module/device/system that receives output 9394. In an embodiment, reporting circuitry 9336 generates a human-readable report containing the results of the analysis (e.g., as indicated by system specification 9356 or other configuration parameters) and reports the human-readable analysis to one or more Can be transmitted to a client device.

112 illustrates governance chips 9400, one or more of which may be used to perform one or more governance functions as described herein. Chip 9400 may be used by any value chain entity that complies with various governance standards, including safety, security, quality, regulatory, financial, or other standards. In embodiments, chip 9400 may use artificial intelligence (AI) to perform governance functions on input data from one or more components of a host device, including governance chip 9400 and/or other devices in communication with the host device. and other techniques may be used. In embodiments, governance chip 9400 may be configured to receive and analyze data to determine situations in which governance may be applied, and may be configured to build one or more models to enforce governance, and then govern governance. Trigger actions in response to a violation, reconstruct data to avoid a governance violation, issue commands to one or more devices in communication with governance chip 9400, and/or perform governance actions using governance chip 9400 in other ways. You can use models to enforce rules, constraints, requirements, quality, or other aspects of governance by performing .

In embodiments, chip 9400 may be configured to receive input data containing data sets to which governance standards may be applied. Input data may include data sets that must comply with one or more safety, security, quality, regulatory, financial, or other standards for a particular domain. In embodiments, multiple governance standards may be applied to a single data set. For example, both safety and quality standards may apply to a given data set. Governance standards may apply only to sets of data based on certain conditions, such as the location or other conditions of a particular device communicating with chip 9400, the current state of a module, device, system, or network, or other such conditions.

Accordingly, chip 9400 may initially analyze a particular data set (e.g., a data set received as input 9492) to determine whether one or more governance standards apply, as described in more detail below. can do. Based on a determination that one or more governance standards apply, chip 9400 may then prioritize applicable standards and create and/or verify models that enforce the governance standards. A model may contain one or more flows to check whether data complies with governance standards, perform actions to ensure compliance with governance standards, take corrective action when governance violations occur, and so on. When multiple governance standards are applied, chip 9400 may create a model that reconciles any potential overlap or conflict between the multiple standards. Chip 9400 may verify the model using test data or other strategies, as described in more detail below.

After the model is created and/or verified, chip 9400 may use the model to enforce governance standards. Chip 9400 may use the model to enforce governance standards on one or more received data sets, including data sets that are not received until after the model has been created and validated. In embodiments, chip 9400 may continuously optimize the model over time to ensure governance compliance as conditions change, to allow review of governance enforcement, and/or to allow other devices to enforce governance enforcement. It can generate reports and other output for you to perform.

In embodiments, chip(s) 9400 may be modular component(s) that can be integrated with a host system in a variety of ways. For example, the chip(s) may be integrated with a mobile host system, a stationary host system, or any other host system that receives input data according to governance. To facilitate such modularity, chip(s) 9400 may be provided partially or fully within a housing (not shown) and may include electrical connectors, optical connectors, and/or wireless connectors (e.g., antennas, Induction coil, etc.) may receive an input 9492 and/or provide an output 9494. Additionally or alternatively, chip(s) 9400 may be integrated with other circuits, processors, systems, etc. on one or multiple substrates/chips.

Chip(s) 9400 may include one or more system-on-chip (SOC), integrated circuit (IC), application specific integrated circuit (ASIC) components to provide the functionality and/or any other functionality ascribed to chip 9400. It may be and/or may include the like. For example, chip 9400 may be provided as part of a SOC that also provides other functionality described in this application. Generally, the components of chip 9400 may include one or more general purpose processing chips configured using software instructions or other code and/or special purpose processing chips (e.g., For example, ASIC) may be included.

Multiple chip(s) 9400 may be used to perform the functions described in this application. For example, multiple chip(s) 9400 may be used to perform analytics and/or governance functions more quickly, or to transfer more complex computations from one chip 9400 to another chip 9400 with a better power source. Serial, parallel, and/or other processing techniques may be used to offload analytics and/or governance functions more efficiently. As another example, one chip 9400 may be used to provide a first analytics and governance function, while another chip 9400 may be used to provide a second analytics and governance function for the same data set. .

In an embodiment, physical input interface 9402 receives one or more inputs 9492 to governance chip 9400 as described herein. Input 9492 may be connected to physical input interface 9402 by other chips, circuits, modules, and/or other components of the host system, or by other devices in communication with the host system (e.g., over a communications network). can be sent. For example, input data may come from a sensor, sensor processing chip/module/circuit, antenna, storage device, network interface, or any other data source for chip(s) 9400 as described herein. You can. Physical input interface 9402 may be coupled to the source(s) of input 9492 via a wired or wireless connection. Input 9492 may include any type of data to which governance may be applied. Input 9492 may also include data that may be stored in repository 9450, such as governance rules/configurations for governance library 9452, one or more digital twins for digital twin library 9454, one or more system specification(s), ) (9456), and/or one or more intelligence module(s) (9458).

Output data 9494 sent from physical output interface 9404 may include report(s) indicating the status of governance functions (e.g., governance compliance and/or violations that may have occurred) (e.g., of a model validation process). (as part of it) may contain data representing the functionality of the generated model, instructions for other modules/devices/systems to enforce compliance with governance standards, etc. In embodiments, the output of chip 9400 may be transmitted by physical output interface 9404 to other chips, circuits, modules, and/or other components of a host system or other devices in communication with the host system, as described herein. can be sent. Physical output interface 9404 may be connected to these components via wired or wireless connections.

In embodiments, chip 9400 may include one or more of a governance analysis module 9410, a governance framework module 9420, and/or a governance output module 9430. In embodiments, governance analysis module 9410 may include circuitry 9412-9416 for receiving and processing input 9492 to determine governance applicability and format input data for application of governance. . Additionally or alternatively, chip 9400 may include a governance framework module 9420 that includes circuitry 9422, 9426 for prioritizing governance, creating governance models, and validating governance models. there is. Additionally or alternatively, chip 9400 may include a governance output module that includes circuitry 9432, 9436 for executing, monitoring, and otherwise processing governance models, optimizing models, and formatting results for output. It may include (9430). The functionality of the various circuits of modules 9410, 9420, and/or 9430 are described in greater detail below.

Processing core(s) 9406 may be one or more processing cores that may be configured to perform any of the functions attributed to chip 9400, with or without assistance from various modules 9410, 9420, and/or 9430. May include (s). For example, processing core(s) 9406 may utilize and/or call various modules to perform various functions described herein. Processing core(s) 9406 may include general-purpose and/or special-purpose processors. In embodiments, processing core(s) 9406 may use serial, parallel, and/or other processing techniques to achieve the functionality described herein.

Accordingly, processing core(s) 9406 may perform functions in addition to those provided by various modules 9410, 9420, and/or 9430. For example, processing core(s) may receive the output of one module (e.g., data extracted by a data set analyzed by governance analysis module 9410) and transfer it to another module (e.g., governance analysis module 9410). It may be provided as input to the framework module 9420 and/or the governance output module 9430. Processing core(s) 9406 may also process the output of any of the module(s) to convert the output to a different format.

In embodiments, processing core(s) 9406 may further operate to store and/or retrieve data to/from storage 9450. For example, processing core(s) 9406 may store and retrieve governance configuration/data in governance library 9452 and/or store and retrieve digital twins within digital twin library 9454 and/or , may store and retrieve system specifications 9456 and/or may store and retrieve intelligence module(s) 9458 for implementing various functions described in this application. In an embodiment, the processing core(s) may use intelligence modules 9458 (which may include one or more of the artificial intelligence modules 8804 of FIG. 104) to provide intelligent services (as described with respect to FIG. 104). Any of the functions of (8800) can be implemented.

Governance analysis module 9410 may receive and process input data 9492 to determine whether governance can be applied and what type of governance can be applied. In embodiments, input data analysis circuitry 9412 may analyze input 9492 to detect conditions indicating that governance applies. For example, the input data may be associated with specific governance requirements (e.g., governance requirements established by the owner of the property corresponding to the location, governance requirements established by a specific state or other governmental entity corresponding to the location, etc.). Can indicate location. As another example, input data may include specific data fields, and one or more values of the data fields may indicate that governance applies. Additionally or alternatively, the input data analysis circuit may access the governance library 9452, digital twin 9454, and/or system specification 9456 to determine whether one or more governance standards apply. For example, governance library 9452 may indicate one or more conditions under which a governance standard applies, which may indicate that a particular governance standard always applies and/or provide other rules, triggers, or conditions indicating that a governance standard applies. . In an embodiment, a digital twin may indicate that input data 9492 relates to a device having a specific state within the digital twin, and the specific state may be associated with a specific set of governance standards. Similarly, system specifications 9456 may provide information about the system corresponding to the data and may indicate if/when governance is applied to the system. Accordingly, using one or more strategies that include analyzing input data 9492 and/or data within storage 9450, chip 9400 may or may not apply governance to input data received as input 9492. You can decide not to.

In embodiments, governance selection circuitry 9414 may determine which of the identified governance requirements apply. For example, one or more governance rules related to safety, security, quality, regulatory, financial, or other standards, the location or other conditions corresponding to the input data, the type of data received as input data, and the type of data received as input data. It may be applied based on various conditions as described above, such as above values, data stored in storage 9450, etc. One or more conditions, triggers, values, or other indications to which a governance requirement applies, as detected by input data analysis circuitry 9412, may each be associated with one or more governance requirements that governance selection circuitry 9414 may retrieve and select. We can respond. In some cases, governance selection circuitry 9414 may need to further analyze the data (e.g., using intelligence module 9458) to determine what governance requirements apply. For example, governance selection circuitry 9414 may process input 9492 using a neural network or other machine learning model to generate a prediction and then determine which governance requirements apply based on the prediction. . In embodiments, multiple intelligence module(s) 9458 may be used to provide various types of AI analytics for governance selections. In an embodiment, intelligence module 9458 may include one or more of artificial intelligence modules 8804 of FIG. 104 .

In embodiments, data analysis circuitry 9416 may perform data analysis to determine and/or extract data to apply governance. For example, data analysis circuitry 9416 may parse or otherwise analyze input 9492 to extract specific values to which governance applies and/or detect specific values to which governance does not apply. In embodiments, data analysis circuitry 9416 may create one or more data structures containing extracted data and format the data structures so that governance standards can be created and/or enforced using the data structures. Data analysis circuitry 9416 may access any of the data stored in repository 9450, which may specify how to detect data values subject to governance for governance requirements selected by governance selection circuitry 9414. You can.

Governance framework module 9420 may receive one or more selected governance requirements from governance analysis module 9410 and develop and validate a model for applying the governance requirements to the data set. In embodiments, prioritization circuitry 9422 may manage multiple and/or overlapping governance requirements by prioritizing governance requirements, resolving conflicts between governance requirements, etc. Prioritization circuitry 9422 may execute one or more prioritization rules contained in governance library 9452 (e.g., by retrieving from governance library 9452 the assigned priority associated with each governance requirement). A priority may be assigned to each of the governance requirements selected by the governance selection circuit 9414 (by using, etc.). In embodiments, prioritization circuitry 9422 may detect whether any of the selected governance requirements overlap or conflict. In some cases, governance requirements may overlap without causing conflict, for example, when a first governance requirement requires a certain minimum standard and a second governance requirement requires a higher standard. In such cases, prioritization circuit 9422 may determine that a higher standard should be used to satisfy both sets of governance requirements. In other cases, such as when governance requirements conflict, prioritization circuitry 9422 may decide to use one or the other conflicting standard based on the priority assigned to each governance requirement.

In embodiments, modeling circuitry 9424 may generate models based on prioritized governance requirements as determined by prioritization circuitry 9422. For example, if the highest priority governance requirement is a set of safety requirements, then the generated model can initially check for safety violations or apply other safety governance requirements. Then, if the second highest priority governance requirement is a set of regulatory governance requirements, the model can enforce regulatory governance after implementing safety governance. In some cases (e.g., due to conflicts), the model may omit certain governance requirements from the model (e.g., quality requirements that conflict with safety requirements). In this way, modeling circuitry 9424 can create a model that specifies flows for enforcing governance over the data set. Modeling circuitry 9424 causes the generated model to reference various digital twins from the digital twin library 9454 specifying information about one or more environments, networks, systems, etc., to retrieve various data that may be needed for inspection and enforcement. You can do it.

In an embodiment, verification circuitry 9426 may verify the generated model, for example, by testing the generated model against test data provided by governance library 9452. In some cases, the selected governance standard may require specific verifications (e.g. verification that the model complies with safety requirements when processing data), so the governance library can successfully fulfill the governance requirement(s) to which the model corresponds. May include test data and/or target output(s) to verify compliance. Additionally or alternatively, verification circuitry 9426 may test the generated model against the digital twin to simulate its effects on one or more devices, networks, systems, etc. In some cases, the simulated effects on the digital twin may be provided as output 9494 (e.g., for analysis/approval on other devices) prior to placing the generated model into governance output module 9430. .

In embodiments, governance output module 9430 may process one or more inputs 9492 using the generated model to enforce governance standards, optimize the model based on various conditions, and/or other It may output processed input, reports, and/or messages to communicate with the device. Model processing circuitry 9432 may receive input 9492 (e.g., inputs analyzed by governance analysis module 9410 as well as input received after the governance model has been created by governance framework module 9420). These can be continuously processed as they are received, so that once deployed, the governance model can be used for new inputs. Model processing circuitry 9432 may use the model to monitor input 9492 and enforce governance standards as specified by the model. For example, model processing circuitry 9432 can generate warnings and alarms, shut down or otherwise modify systems (e.g., if safety parameters are exceeded), and modify/transform/transform data to comply with governance. You can configure it, etc. In embodiments, chip 9400 may transmit messages and/or commands to other devices and systems to enforce governance requirements. In these cases, model optimization circuitry 9434 may cause output and reporting circuitry 9436 to send such messages and/or instructions, as described in more detail below.

In embodiments, model optimization circuitry 9434 may perform live optimization of the governance framework/model by continuously monitoring various input conditions and data. For example, in response to a change in location or some other condition, a different set of governance requirements may begin to apply. Model optimization circuitry 9434 allows governance framework module 9420 to prioritize new governance requirements, update the model accordingly, and/or validate the updated model, as described above. Different sets of governance requirements can be implemented by allowing them to be recreated and/or modified. Additionally or alternatively, model optimization circuitry 9434 may continuously verify the output of model processing circuitry 9432 to ensure that the model used by model processing circuitry 9432 is performing properly. As described above, model optimization circuitry 9434 may perform verification by referring to verification data/requirements stored in storage 9450.

In embodiments, output and reporting circuitry 9436 may send output including data to be processed by model processing circuitry 9432 as well as messages and/or instructions to be transmitted to other modules, devices, systems, etc. Accordingly, chip 9400 may enforce governance requirements by causing other devices to change states (e.g., turn off/on) or otherwise perform governance actions. Additionally or alternatively, output and reporting circuitry 9436 may be configured to report, including the results of verification, alerts, or other non-compliance with governance, for review/analysis by other chips, modules, systems, or devices. , reports indicating governance conflicts, etc. can be generated. Output and reporting circuitry 9436 can cause any output to be sent as output 9494.

113 illustrates prediction, classification, and recommendation chips 9500, one or more of which may be used to perform one or more prediction, classification, and/or recommendation functions as described herein. Chip 9500 may be used by any value chain entity to perform prediction, classification, and/or recommendation. In embodiments, chip 9500 may perform prediction, classification, and/or recommendation functions on input data from one or more components of a host device, including chip 9500 and/or other devices in communication with the host device. Artificial intelligence (AI) and other technologies may be used to do this. In embodiments, chip 9500 may analyze and classify incoming data according to a given set of specifications, develop and/or provide optimized predictive models according to a given set of specifications, and/or provide data according to a given set of specifications. It may be configured to provide recommended actions based on classification and predictive modeling.

In embodiments, chip 9500 may be configured to store any data including media data such as image/video/audio data, datasets including transaction data, biometric data, motion capture data, pathology data, and/or other such data. It may be configured to receive input of various types and analyze such data to determine additional information (e.g., metadata) about the input data, objects or entities appearing in the input data, etc. Chip 9500 may then classify the input, objects or entities appearing in the input, etc. using various classification techniques, as described in detail below. Chip 9500 may output the classification as output 9594 for use by other modules, devices, systems, etc.

In embodiments, chip 9500 may develop one or more conditions for use in generating a predictive model. Conditions can be developed based on classification. That is, based on classifying a particular object, entity, or grouping thereof, one or more of the Conditions may be developed and selected for predictive analysis. Chip 9500 may then create and utilize a predictive model to predict the effect of an action involving an object, entity, or grouping thereof, based on the updated data, as described in more detail below. The prediction model can be further optimized.

In an embodiment, chip 9500 may use various system specifications to determine the scope of actions that may be taken with respect to various entities, including: one or more actions, one or more direct or indirect objects or other entities on which actions may be taken; You can create an action matrix containing one or more action modifiers, etc. Chip 9500 then analyzes and determines which action(s) should be taken from the action matrix (e.g., using the generated predictive model) and transmits an output 9594 that results in performance of the selected action. , may provide feedback to improve the functionality of the classification, prediction, and recommendation functions, as described in more detail below.

In embodiments, chip(s) 9500 may be modular component(s) that can be integrated with a host system in a variety of ways. For example, the chip(s) may be integrated with a mobile host system, a stationary host system, or any other host system that receives input data for prediction, classification, and/or recommendation tasks. To facilitate such modularity, chip(s) 9500 may be provided partially or completely within a housing (not shown) and may include electrical connectors, optical connectors, and/or wireless connectors (e.g., antennas, Induction coil, etc.) may receive an input 9592 and/or provide an output 9594. Additionally or alternatively, chip(s) 9500 may be integrated with other circuits, processors, systems, etc. on one or multiple substrates/chips.

Chip(s) 9500 may include one or more system-on-chip (SOC), integrated circuit (IC), application specific integrated circuit (ASIC) components to provide the functionality and/or any other functionality ascribed to chip 9500. It may be and/or may include the like. For example, chip 9500 may be provided as part of a SOC that also provides other functionality described in this application. Generally, the components of chip 9500 may include one or more general purpose processing chips configured using software instructions or other code and/or special purpose processing chips (e.g. For example, ASIC) may be included.

Multiple chip(s) 9500 may be used to perform the functions described in this application. For example, multiple chip(s) 9500 can perform AI-assisted functions faster, offloading more complex computations from one chip 9500 to another chip 9500 with a better power source. Thus, serial, parallel, and/or other processing techniques can be used to perform AI-assisted functions more efficiently. As another example, one chip 9500 may be used to provide a first AI-assisted functionality described herein, while another chip 9500 may provide a second AI-assisted function based on the same input 9592. It can be used to provide functionality.

In an embodiment, physical input interface 9502 receives one or more inputs 9592 to chip 9500 as described herein. Input 9592 may be connected to physical input interface 9502 by other chips, circuits, modules, and/or other components of the host system, or by other devices in communication with the host system (e.g., over a communications network). can be sent. For example, input data may come from a sensor, sensor processing chip/module/circuit, antenna, storage device, network interface, or any other data source for chip(s) 9500 as described herein. You can. Physical input interface 9502 may be coupled to the source(s) of input 9592 via a wired or wireless connection. Input 9592 may include any type of data to which governance may be applied. Input 9592 may also include data that may be stored in storage 9550, such as governance rules/configurations for governance library 9552, one or more digital twins for digital twin library 9554, one or more system specification(s), ) (9556), and/or one or more intelligence module(s) (9558).

Output data 9594 transmitted from physical output interface 9504 may include one or more classified, predicted, and/or recommended actions, as well as information about the input to chip 9500, generated by chip 9500. It may include one or more reports to provide data, functions of the chip 9500, etc.

In embodiments, chip 9500 may include one or more of a classification module 9510, a prediction module 9520, and/or a recommendation module 9530. In an embodiment, classification module 9510 may include circuitry 9512-9516 to receive and extract/separate data, analyze data, and classify data. Additionally or alternatively, chip 9500 may develop and/or otherwise utilize conditions (e.g., based on classifications provided by classification module 9510) and generate predictions using prediction models. and may include a prediction module 9520 including circuits 9522 and 9526 for optimizing the prediction model. Additionally or alternatively, chip 9500 may generate a recommended action matrix, analyze the decision criteria applied (e.g., to select one or more recommended actions), view the recommended action(s), and or otherwise perform, and may include a recommendation module 9530 that includes circuitry 9532, 9538 to provide feedback data for use by various modules and/or circuits of chip 9500. The functionality of the various circuits of modules 9510, 9520, and/or 9530 are described in greater detail below.

Processing core(s) 9506 may be one or more processing cores that may be configured to perform any of the functions attributed to chip 9500, with or without assistance from various modules 9510, 9520, and/or 9530. May include (s). For example, processing core(s) 9506 may utilize and/or call various modules to perform various functions described herein. Processing core(s) 9506 may include general-purpose and/or special-purpose processors. In embodiments, processing core(s) 9506 may use serial, parallel, and/or other processing techniques to achieve the functionality described herein.

Accordingly, processing core(s) 9506 may perform functions in addition to those provided by various modules 9510, 9520, and/or 9530. For example, processing core(s) may receive the output of one module (e.g., classification data generated by classification module 9510) and process it into another module (e.g., prediction module 9520). It can be provided as input. Processing core(s) 9506 may also process the output of any of the module(s) to convert the output to a different format.

In embodiments, processing core(s) 9506 may further operate to store and/or retrieve data to/from storage 9550. For example, processing core(s) 9506 may store and retrieve governance configurations/data in governance library 9552 and/or store and retrieve digital twins in digital twin library 9554 and/or , may store and retrieve system specifications 9556 and/or may store and retrieve intelligence module(s) 9558 for implementing various AI-assisted functions described in this application. In an embodiment, the processing core(s) may use intelligence modules 9558 (which may include one or more of the artificial intelligence modules 8804 of FIG. 104) to provide intelligent services (as described with respect to FIG. 104). Any of the functions of (8800) can be implemented.

The classification module 9510 may receive input data, isolate/extract the input data, analyze the data, and classify the data. In embodiments, data isolation circuitry 9512 may receive input data 9592 and extract or otherwise isolate the input data prior to analysis. For example, input data 9592 may be one or more data streams or data sets including image/video data, transaction data, biometric data, diagnostic data, or any other type of data as described herein. You can. Data isolation circuitry 9512 may isolate such data from the data stream/set (e.g., by identifying data for analysis, extracting data, converting/reformatting data, etc.). For example, data isolation circuitry 9512 may extract images from video, convert dialogue to text, extract relevant data from larger data sets, etc.

In embodiments, analysis circuitry 9514 may analyze isolated data and/or other data to determine information for classification. For example, the analysis circuitry may perform image analysis on the image to identify one or more objects appearing in the image and transaction metadata (e.g., sender/recipient identity, type of transaction, etc.) Transaction data may be analyzed to determine personal information, biometric data may be analyzed to determine personal metadata (e.g., identity, demographic information, etc.), /or analyze motion seen in video data (e.g., to determine response information), analyze diagnostic data (e.g., to determine abnormalities or other conditions from a diagnostic data set), etc. In some cases, analysis circuitry 9514 may utilize information stored in storage 9550 to perform analysis. For example, analysis circuitry 9514 may use various digital twins from digital twin library 9554 and/or system specifications 9556 to obtain information about various systems corresponding to input data (e.g. Intelligence module 9558 may be used to perform various analyzes (e.g., provide additional information about the device or other entity corresponding to the input data, allow interpretation of the input data, etc.) machine vision intelligence module that performs recognition), etc. Additionally or alternatively, analysis circuitry 9514 may structure data for classification by classification circuitry 9516.

In embodiments, classification circuitry 9516 may perform classification tasks on isolated data and/or any additional data generated by analysis circuitry 9514. Classification circuit 9516 may use one or more machine learning or other AI-assisted techniques (e.g., regression, naive Bayes, stochastic gradient descent, k-nearest neighbors, decision-making) to classify the data. tree, random forest, etc.) can be used. For example, classification circuitry 9516 may classify objects appearing in an image (e.g., by identifying types of different groupings of objects) and whether a transaction is abnormal (e.g., by type of transaction). /classify transaction data (e.g., by whether suspicious, etc., by type of party to the transaction, etc.), and classify people according to biometric data (e.g., by demographics, by type of emotion, etc.) may classify motion data (e.g., by response type), may classify diagnostic data (e.g., to identify pathology or other abnormalities in individual or population data), etc. am. Classification circuitry 9516 may utilize unsupervised machine learning techniques to group data isolated by data isolation circuitry 9512 and/or generated by analysis circuitry 9514 and/or supervised learning for specific tasks. Technologies (e.g., trained models that may be stored in repository 9550 as intelligence modules 9558) may be used. Accordingly, chip 9500 can be configured for a particular classification task by storing appropriate configuration data (e.g., trained models) in storage 9550.

Prediction module 9520 may develop, utilize, and/or optimize a prediction model to generate predictions based on data received as input 9592 and/or one or more specifications 9556. In embodiments, condition development circuitry 9522 may develop conditions that can be used to generate a prediction model based on the classification performed by the classification model. When classification circuitry 9516 detects one or more classifications, condition development circuitry 9522 may use the predictive model to select one or more conditions associated with the classification to target. For example, based on classification circuitry 9516 recognizing a particular type of object within an image, condition development circuitry 9522 may determine target variables associated with the detected type of object for use in the development of a predictive model (e.g. For example, number/quantity/frequency of objects or other target variables that are a function of objects) can be developed. As another example, based on classification circuitry 9516 recognizing a particular type of transaction in the transaction data, condition development circuitry 9522 may select an estimate of future transactions of the detected type as a target variable. As another example, based on classification circuitry 9516 recognizing a particular type of behavior or demographics, condition development circuitry 9522 may perform an evaluation of object or group behavior (e.g., based on unsafe behavior). Target variables can be developed including security estimates, cognitive assessments, etc. As another example, based on the classification circuit 9516 recognizing a particular type of pathology, the condition development circuit 9522 may develop target variables including the estimated prevalence of the pathology, population changes, cost of addressing the pathology, etc. You can. In some cases, one or more stored system specifications 9556 may indicate which conditions are available for targeting and/or should be targeted. Accordingly, chip 9500 may be configured for a specific system/task/domain by storing specific system specifications 9556.

In embodiments, predictive modeling circuitry 9524 may create a predictive model for predicting a target variable based on input data, data generated by analysis circuitry 9514, and/or classification data generated by classification circuitry 9516. The target variable generated by the condition development circuit 9522 can be used to train. That is, predictive modeling circuitry 9524 can train a model to predict a target variable using a training data set that includes any of the data described above. Predictive modeling circuitry 9524 may use various AI-assisted learning techniques (e.g., neural networks, deep learning, etc.) to develop a model based on selected target variables.

Additionally or alternatively, predictive modeling circuitry 9524 may utilize predictive models to generate predictions based on various modeling inputs. Modeling input may include input 9592 (e.g., input data used by classification module 9510 and/or a new set of input data as described above), isolated data generated by data isolation circuit 9512, and /can be derived from extracted input data, data generated by the analysis circuit 9514, classification generated by the classification 9516, etc. That is, any of the data received as input 9592 and/or generated by chip 9500 may be used as input to the prediction model. Predictive modeling circuitry 9524 may provide various inputs to a prediction model to generate a prediction, including one or more discrete and/or continuous values (e.g., predicted scores and/or classification), one or more confidence levels, and It may include etc.

In embodiments, prediction model optimization circuitry 9526 may optimize a prediction model by updating a training data set, retraining the prediction model, selecting different target variables, developing a new model, etc. For example, a predictive model may periodically update a training data set, received as input 9592 and/or applied to any of analysis circuitry 9514, classification circuitry 9516, and/or predictive modeling circuitry 9524. The model can be retrained using new data generated by Additionally or alternatively, predictive model optimization circuitry 9526 may monitor the accuracy of predictions by monitoring input data 9592 from digital twin library 9554 and/or one or more digital twin(s) over time. You can. For example, if predictive modeling circuitry 9524 repeatedly predicts future conditions with high confidence, but predictive model optimization circuitry 9526 later determines that the predicted condition will not occur, predictive model optimization circuitry 9526 may This may result in updating/modifying the training data set and/or training parameters and retraining the prediction model to provide more accurate predictions.

In embodiments, recommendation module 9530 may provide recommendations based on various specifications 9556, classifications generated by classification module 9510, and/or predictions generated by prediction module 9520. In embodiments, action matrix circuitry 9532 may generate a matrix (e.g., an N-dimensional array that may contain a simple list) of potential actions that can be taken with respect to a particular task, system, or domain. . For example, system specification 9556 may provide a first set of potential actions, a second set of potential entities on which actions may be taken, a third set of modifiers for the actions, etc., and thus an action matrix circuit. 9532 may generate a matrix of potential actions that can be recommended. Additionally or alternatively, certain actions, entities, etc. may be associated with input data, analysis performed by analysis circuitry 9514, classification generated by classification circuitry 9516, and/or predictive modeling circuitry 9524. Based on predictions, they can be automatically identified and added to the action matrix. Additionally or alternatively, certain actions, entities, etc. may be automatically maintained or removed from the action matrix (e.g., from governance library 9552) based on governance data. Actions may include instructions addressed to digital and/or real-world entities, such as instructions to be performed by a human, computing device, system, module, etc.

In embodiments, decision analysis circuitry 9534 may analyze some or all of the actions in the action matrix to determine one or more recommended actions. Decision analysis circuitry 9534 determines the effects of a particular action (which may involve using, for example, prediction module 9520, intelligence module 9558, and/or some other resource to predict the effects of an action). You can utilize the digital twin(s) in the digital twin library 9554 to simulate. Additionally or alternatively, decision analysis circuitry 9534 may use one or more governance requirements stored in governance library 9552 to determine whether a particular action violates a governance requirement (e.g., because it is unsafe or illegal) and/or You may decide that certain actions are required to comply with governance requirements. As a first example, detecting a particular type of object that appears in one or more images (e.g., as determined by classification module 9510) and detecting an object (e.g., as determined by prediction module 9520) Based on predicting that the target variable may decrease, the decision analysis circuit 9534 may interact with the object to increase the target variable (e.g., by moving the object or otherwise interacting with the object). I can recommend doing this. As another example, detecting a particular type of transaction from transaction data (e.g., as determined by classification module 9510) and determining if the transaction has a specific negative Based on what it predicts may result in the outcome, decision analysis circuitry 9534 may recommend preventing similar transactions in the future. As a third example, detecting a particular type of condition from biometric or diagnostic data (e.g., as determined by classification module 9510) and (e.g., as determined by prediction module 9520) Based on predicting that a particular pathology or other condition is present, decision analysis circuitry 9534 may recommend a particular intervention. As a fourth example, detecting a particular type of individual and/or group behavior (e.g., as determined by classification module 9510) and conditions (e.g., as determined by prediction module 9520) Based on predicting that this is becoming abnormal or unsafe, decision analysis circuitry 9534 may recommend shutting down a particular location, system, or taking other corrective action.

In embodiments, recommended action and reporting circuitry 9536 may perform one or more recommended actions and/or cause other module(s), device(s), system(s), etc. to perform recommended actions. may cause transmission of an output message (e.g., via output 9594). Additionally or alternatively, recommended action and reporting circuitry 9536 may generate reports that may include classifications, predictions, recommendations, and/or any of other data received or generated by chip 9500. there is. Recommended action and reporting circuitry 9536 may send reports as output 9594 to other modules, devices, systems, etc.

In embodiments, feedback circuitry 9538 monitors results associated with classifications, predictions, and/or recommended actions to determine whether the classifications and/or predictions were accurate, whether the recommended actions had the desired/predicted impact, etc. can do. Accordingly, the feedback circuit 9538 may utilize one or more digital twin(s) within the digital twin library 9554 to monitor one or more devices, systems, environments, etc. In such embodiments, the digital twin(s) are continuously monitored by other components (e.g., as described elsewhere in this application) that keep the digital twin updated for monitoring by feedback circuitry 9538. can be updated. Based on the monitored results, feedback circuit 9538 adjusts (e.g., retrains) any models used by classification module 9510, prediction module 9520, and/or recommendation module 9530. can do.

Accordingly, classification, prediction, and recommendation chip 9500 may perform a wide variety of classification, prediction, and recommendation tasks, including any of the classification, prediction, and recommendation tasks described herein, but some examples include classification , may be useful in demonstrating the flexibility and functionality of the prediction, and recommendation chip 9500. According to a first example, chip 9500 automatically analyzes and classifies satellite images (e.g., recognizes specific vegetation types, densities and locations, animal populations and movements, etc.) and based on classified objects within the images. provide forecasts (e.g. crop evaluation, fire risk assessment, water allocation and pricing, etc.) and make recommendations based on classification and forecasts (e.g. crop production adjustments, brush clearing, increase in insurance reserves, water allocation) It can be configured to provide a reduction of, etc.). According to this first example, enterprise resource planning system inputs (e.g. inventory, pricing, accounting, sales, employee information), customer relationship management system inputs (e.g. customer data, payment methods, etc.), security system inputs A wide variety of inputs 9592 may be used, including inputs including (e.g., data access and management, surveillance video, authentication data), crime statistics, police reports, cost of living reports, etc. Additionally, system specifications 9556 in this example may indicate that various actions may include increasing/decreasing/maintaining storage times, products or services offered, adjusting security levels, etc. Moreover, system specification 9556 can include a list of stores, products, or services that can be reconciled, so that a three-dimensional action matrix can be developed representing the actions, stores, and reconciliations. According to a second example, chip 9500 automatically analyzes and classifies financial transactions (e.g., to recognize fraud or theft, type of purchase, contract, customer, product, etc.) and transaction data (e.g. For example, providing forecasts based on demand response, fraud estimation and response, asset allocation, etc.), classifying and forecasting (e.g., increasing production, reallocating inventory, investing in security and enforcement, adjusting profit projections, reallocating assets). etc.) may be configured to provide recommendations based on. According to a third example, the chip 9500 automatically analyzes and classifies biometrics (e.g., recognizes faces, voices, gestures, identifies groups, evaluates emotions, etc.) and biometric data ( configured to provide predictions based on (e.g., individual or group behavior, security, cognitive assessments, etc.) and provide recommendations based on classifications and predictions (e.g., health or psychological screening, security certification/assessment, etc.) It can be. According to a fourth example, chip 9500 automatically analyzes and classifies motion capture data (e.g., classifies behavior as normal or abnormal, safe or unsafe, etc.) and based on the motion capture data, It may be configured to provide predictions (e.g., group behavior based on individual responses, etc.) and provide recommendations based on classification and predictions (e.g., intervention, rerouting of group flow patterns, etc.). According to a fifth example, the chip 9500 automatically analyzes and classifies pathology data (e.g., detects disease, population health, disease prevalence and spread, etc.), and analyzes and classifies pathology data (e.g., disease outbreak, population change, health care costs, etc.), and may be configured to provide recommendations based on classification and predictions (e.g., quarantine, allocation of medical resources, adjustment of insurance premiums, etc.).

Additive Manufacturing

114-121 illustrate various embodiments of an additive manufacturing platform. In embodiments, the additive manufacturing platform may be a standalone system or may be integrated into a larger system, in which case the additive manufacturing platform is a value chain entity. In embodiments, “additive manufacturing” allows a 3D digital model (CAD design) to be converted into a three-dimensional object by depositing multiple thin layers of material, such as according to a series of two-dimensional cross-sectional deposition maps. Refers to a collection of versatile manufacturing technologies for rapid prototyping and/or manufacturing of parts.

Accordingly, the term “additive manufacturing platform” as used in this application includes a platform that prints, builds, or otherwise produces 3D parts and/or products at least in part using additive manufacturing technologies. Additive manufacturing platforms include technologies such as 3D printing, vapor deposition, polymer (or other material) coating, epitaxial and/or crystalline growth approaches, and other technologies, alone or in combination with other technologies, such as cutting or assembly techniques. This makes it possible to manufacture a three-dimensional product from a design through a process that forms successive layers of the product, reaching a finished component or system with optional intermediate or subsequent steps. The design may be in the form of a data source, such as a computer-aided design package or an electronic 3D model created through a 3D scanner. 3D printing or other additive manufacturing processes then involve forming a first material layer and then adding successive material layers, each new material layer following the previous one until the entire designed three-dimensional product is complete. is added on the formed layer of material. The additive manufacturing platform may be a stand-alone unit, a sub-unit of a larger system or production line, and/or may incorporate other non-additive manufacturing features, such as subtractive manufacturing features, pick-and-place features, It may include coating features, finishing features (e.g., etching, lithography, painting, polishing, etc.), two-dimensional printing features, etc. Additionally, the platform includes three-dimensional additive manufacturing machines configured for rapid prototyping, three-dimensional printing, two-dimensional printing, freeform manufacturing, solid freeform manufacturing, and stereo-orithography; Subtractive manufacturing machines, including computer numerical control manufacturing machines; It may include injection molding machines, etc.

114 is a schematic diagram illustrating an example environment of an autonomous additive manufacturing platform 10110 according to some embodiments of the present disclosure. The platform operates within manufacturing nodes 10100, which in turn are part of a larger network of value chain entities. Fabrication node 10100 may be made of metallic materials, biocompatible materials, bioactive materials, biological materials, or other more conventional additive manufacturing materials, or as understood in the art or documents incorporated by reference in this application. and an additive manufacturing unit 10102, such as a 3D printer, for printing in other additive manufacturing types described herein. Manufacturing node 10100 may include a pre-processing system 10104, a post-processing system 10106, and a material handling system 10108, among other elements. The Autonomous Additive Manufacturing Platform (10110) automates digital production workflows leading to better results at all stages of operation, specifically from initial design through printing and supply chain logistics to point of sale, service and utilization of the resulting output. and help optimize it. In an embodiment, user interface 10112 receives design and modeling data from design and simulation system 10116 as well as input data from data source 10114. The data processing and intelligence component 10118 of the autonomous additive manufacturing platform 10110 may utilize, e.g., machine learning or other algorithms, neural networks, etc., to process input data and calculate an optimal set of process parameters for printing or other additive manufacturing. Implement artificial intelligence systems involving expert systems, models, etc. The process control component 10120 of the autonomous additive manufacturing platform 10110 then adjusts one or more process parameters in real time, and the additive manufacturing unit 10102 uses these process parameters to complete the additive manufacturing process. In embodiments, a finishing system 10121 at a manufacturing node 10128, such as a cutting system, assembly system, further processing system, etc., performs further processing in an optionally iterative sequence with a lamination stage to form a finished item (e.g. For example, parts, components, or finished products). In embodiments, the resulting product may then be optionally packaged in a packaging system 10122 and shipped using one or more value chain network (VCN) entities 10126 and a delivery system 10124 directly to the end customer. . In other embodiments, the additive manufacturing platform 10110 and/or the set of additive manufacturing units 10102 may be a handheld unit, a robot, or other unit equipped with autonomous mobility, and/or a vehicle, including general purpose and special purpose vehicles. It may include a portable or otherwise mobile unit, such as a unit positioned within or on the unit. In such cases, actions from design to delivery may occur in parallel with the mobility of the units 10102 and in coordination with the location and mobility of other value chain network entities 10126 by the additive manufacturing platform 10110. In one of many possible examples, a set of autonomous mobile 3D printing units could be coordinated to a point of service operation, such as a set of home or business locations, where they may produce tools, parts, or It may be configured to print other items. In embodiments, additive manufacturing, including design creation, design review, preprocessing, and printing steps, may begin while unit 10102 is in transit to a service point. In another example, a mobile autonomous additive manufacturing unit 10102 (autonomous, semi-autonomous, or with an operator) and a packaging unit may, for example, create customization elements (e.g., a final coating in a selected color, customer-specific design elements, etc.) on the go. By adding and optionally completing final packaging on the go, the final steps of manufacturing can be completed on the go. In embodiments, one or more components of the additive manufacturing platform 10110 may be deployed in or integrated with a smart container or smart package, as described elsewhere in this application and in documents incorporated by reference herein. . In embodiments, a set of additive manufacturing units 10102 may be integrated within or with a set of robotic systems, such as mobile and/or autonomous robotic systems. For example, the additive manufacturing unit 10102 may be included within the housing or body of a robotic system, such as a multipurpose/general purpose robotic system such as one that simulates human or other animal species capabilities. Alternatively or additionally, the additive manufacturing unit 10102 may be configured to deliver additive deposition from a nozzle disposed on a working end of a robot arm or other assembly. In embodiments, multiple additive manufacturing units 10102, or multiple nozzles, printheads, or other operational elements may be integrated with a single mobile, autonomous, and/or multi-purpose robotic system, e.g., wherein one additive manufacturing A unit 10102 is housed within the body of the robotic system (e.g., within a chamber such as a vacuum chamber, pressurization chamber, heating chamber, etc.) and prints the layer, and other additive manufacturing units 10102 may print the layer, e.g., on an arm or similar element of the robot. A layer is printed or otherwise operated on an external site, such as a target location of a machine, product, etc., by nozzles, printheads, etc. placed on it. In an embodiment, multiple printing/layering elements are served by a common material source, such as a thermoplastic material. In embodiments, multiple material sources are available for internal and external printing/lamination elements. In embodiments, the internal printing elements operate within chambers using materials that require control over the printing environment, or for expensive production elements such as parts intended for long-term use, such as metal fabrication parts. In embodiments, an external work unit may perform tasks that use materials or require other materials and/or produce disposable tools, grips, supports, fasteners, etc., specifically to support operations such as repair or replacement operations. It has the same different purpose. In embodiments, an external printing/laminating unit may be coupled with a robotic arc welding unit to provide a set of printing/laminating steps and a series of arc welding steps in series or parallel to perform operations, e.g., at an external location, work piece, etc. are combined. In embodiments, nozzles, printheads, etc. may optionally also be associated with additive manufacturing elements to encapsulate and/or shield external working units such as temporary chambers, balloons, tents, or other volumes that isolate the area, layer, etc. to be printed. The assembly may be provided to encapsulate or shield a workpiece or target location for printing/lamination within the same shielded/isolated space. In embodiments, the encapsulated/shielded region may be sealed to allow for pressurizing, depressurizing, creating a vacuum, introducing material for deposition, etc. In embodiments, the encapsulation/shielding may use additively manufactured elements, or combinations thereof with other elements. In embodiments, AI system 10212 may automate one or more of the design, construction, scheduling, coordination, and/or execution of sets of robotic tasks and sets of additive manufacturing tasks, such that the included mobile robots and additive manufacturing units Capabilities are coordinated across a variety of tasks at the right time (e.g., an internal 3D printer or other additive manufacturing unit 10102 can print tools, workpieces, parts, etc. for a later task while the robotic unit performs the current task). ) and/or tasks are coordinated across a fleet or workforce of robotic units, additive manufacturing units, and integrated combinations thereof (e.g., whether units are assigned to tasks depending on location, robotic capabilities, additive manufacturing capabilities, and other factors). if matched).

In embodiments, material handling system 10108 provides storage, movement, control, and handling of materials through manufacturing and distribution processes. For example, material handling system 10108 may feed, orient, load/unload, or otherwise manipulate metallic materials, biocompatible materials, bioactive materials, biological materials, or other more conventional additive manufacturing materials within a manufacturing space. You can. In embodiments, material handling system 10108 may be semi-automatic or fully automated and may include one or more robotic units for material handling.

In embodiments, material handling system 10108 optionally captures material (e.g., unused material from a job and/or re-captures unused material from a job site, e.g., used, damaged material), optionally within the same housing, unit, or system. may include or be integrated with a material capture and processing system 10127 to render material suitable for use as source material, for example, by the following steps: : (a) automatically analyzing an item to determine its compatibility for use as a source material (e.g., by machine vision, chemical testing, image-based testing, weighing of the item, etc.) for a given type of metal; , by identifying it as an alloy, polymer or plastic); (b) cleaning, filtering, disassembling or otherwise pretreating the item or material, such as to remove nonconforming materials; (c) rendering the solid item or material into a thermoplastic state, such as by controlled heating, such as according to a material-specific heating profile; (d) filtering or otherwise treating the material, such as to remove defects; (e) capacity for wider systems for managing operations, for example, including cooling and/or otherwise processing materials into wires, powders, meshes, rods, filaments, etc., until the need for operations arises; and storing the item in an appropriate container or form factor for later use, with appropriate reporting of availability; (f) delivering the item for an additive manufacturing operation; and/or (g) reporting on recapture and reduction actions, including material cost savings, recycling cost savings, and/or time savings. For example, in embodiments, damaged parts may be melted and reprinted in situ. For example, in embodiments, materials that would otherwise be discarded or recycled may become available in the field without requiring reverse logistics. In embodiments, a common heat source is used, along with alternative heating points at different temperatures, to bring the recaptured material into a thermoplastic state and prepare the material for additive manufacturing operations.

Value chain entity 10126 refers to any of the wide variety of assets, systems, devices, machines, components, equipment, facilities, individuals, or other entities mentioned throughout this disclosure or in documents incorporated by reference herein. , various entities involved in the supply, demand, distribution or supply chain environment, such as, but not limited to: machinery and its components (e.g., delivery vehicles, forklifts, conveyors, loading machines, cranes, lifts, transporters, trucks, loading machines, unloading machines, packing machines, picking machines, and many others, including robotic systems such as physical robots, collaborative robots (e.g., “cobots”), drones, autonomous vehicles, software bots, etc. ); Workers (e.g. designers, engineers, process supervisors, supply chain managers, floor managers, demand managers, forwarding workers, delivery workers, barge workers, port workers, dock workers, train workers, ship workers, distribution or fulfillment center workers, warehouse workers) operators, fleet drivers, business managers, marketing managers, inventory managers, freight handlers, inspectors, delivery personnel, environmental control managers, financial asset managers, security personnel, safety personnel, etc.); Suppliers (e.g. suppliers of all types of goods and related services, component suppliers, ingredient suppliers, material suppliers, manufacturers, etc.); Customers (including consumers, licensees, business owners, businesses, value-added and other resellers, retailers, end users, distributors, and others who may purchase, license or otherwise use the category of goods and/or related services) box); Retailers (including, for example, online retailers in the form of e-commerce sites, traditional offline retailers, pop-up stores, etc.); Value chain processes (e.g. shipping process, transport process, maritime process, inspection process, transport process, loading/unloading process, packing/unpacking process, configuration process, assembly process, installation process, quality control process, environmental control process ( (e.g. temperature control, humidity control, pressure control, vibration control, etc.), boundary control processes, port-related processes, software processes (including applications, programs, services, etc.), packing and loading processes, financial processes (e.g. For example, insurance process, reporting process, transaction process, etc.), testing and diagnostic process, security process, safety process, reporting process, asset tracking process, etc.); Wearable and portable devices (e.g. mobile phones, tablets, dedicated handheld devices for value chain applications and processes, data collectors (including mobile data collectors), sensor-based devices, watches, glasses, hearing devices, head-worn devices, clothing integrated devices, arm bands, bracelets, neck-worn devices, AR/VR devices, headphones, etc.); A wide range of operational facilities (e.g. loading and unloading docks, storage and warehousing facilities, vaults, distribution facilities and fulfillment centers, air travel facilities (including aircraft, airports, hangars, runways, refueling depots, etc.), marine facilities (e.g. , port infrastructure facilities (e.g., docks, yards, cranes, roll-on/roll-off facilities, ramps, containers, container handling systems, waterways, locks, etc.), shipyard facilities, floating assets (e.g., ships, barges, boats, etc.), facilities and other items at the point of origin and/or destination, transport facilities (such as container ships, barges and other floating assets, as well as land-based vehicles and other delivery systems used to transport goods, such as trucks, trains, etc.); items or factors (i.e., demand factors) that take into account demand (including market factors, events, etc.); items or factors that take into account supply (i.e., supply factors) (including market factors, weather, components, etc.) and availability of materials, etc.); logistics factors (e.g., availability of travel routes, weather, fuel prices, regulatory factors, (on vehicle, in container, in package, in warehouse, in fulfillment center, on shelf, etc.) availability of space, etc.); routes for transportation (e.g. waterways, roads, air routes, railways, etc.); robotic systems (including mobile robots, cobots, robotic systems to assist human workers, robotic delivery systems, etc.) (e.g., for package delivery, site mapping, monitoring or inspection, etc.); drones (including for package delivery, site mapping, monitoring or inspection, etc.); autonomous vehicles (e.g., for package delivery); software platforms (e.g., enterprise resource planning platforms, customer relationship management platforms, sales and marketing platform, asset management platform, Internet of Things platform, supply chain management platform, platform as a service platform, infrastructure as a service platform, software-based data storage platform, analysis platform, artificial intelligence platform, etc.);

Manufacturing node 10100 may also connect to other nodes, such as manufacturing node 10128, via connectivity facilities to form a distributed manufacturing network 10130. Additionally, there is an additive manufacturing unit (10102), a pre-processing system (10104), a post-processing system (10106), a material handling system (10108), an autonomous additive manufacturing platform (10110), a user interface (10112), a data source (10114), and a design. and simulation system 10116, as well as different systems within manufacturing node 10100, including the different parts and products being printed, may be referred to as distributed manufacturing network entities.

In embodiments, connectivity facilities include network connections (including various configurations, types, and protocols for fixed and wireless connections), Internet of Things devices, edge devices, routers, switches, access points, repeaters, mesh networking systems, interfaces, and ports. , including application programming interfaces (APIs), brokers, services, connectors, wired or wireless communication links, human-accessible interfaces, software interfaces, micro-services, SaaS interfaces, PaaS interfaces, IaaS interfaces, cloud capabilities, etc. Includes various connectivity facilities described throughout the disclosure and documents incorporated by reference herein, whereby data or information may be transferred between systems or subsystems of autonomous additive manufacturing platform 10110, as well as distributed manufacturing network entities. It can be exchanged with other systems, such as cloud-based or on-premises enterprise systems (e.g., accounting systems, resource management systems, CRM systems, supply chain management systems, etc.). In embodiments, connectivity facilities may be configured to enable self-organization or self-organization of connectivity, data storage, computation, data processing, packet routing, data filtering, quality of service, error correction, packet security, session management, etc. , uses, includes, or is integrated with artificial intelligence or autonomous capabilities as described in this application and/or in documents incorporated by reference in this application. In embodiments, the additive manufacturing units 10102 may be optionally configured such that a set of such additive manufacturing units 10102 can operate as a coordinated mesh over a defined network infrastructure (including physical and/or virtual network resources). It can include wireless mesh network nodes, such as RF repeaters, using software-defined bandpass filtering. In embodiments, the additive manufacturing unit 10102 may be configured to control utilization of data paths between the additive manufacturing unit 10102 and other additive manufacturing units 10102 and/or between the additive manufacturing unit 10102 and various edge, cloud, It may include a network coding system to control the utilization of data paths between premises, telecommunication networks and other information technology systems.

Additive manufacturing unit 10102 may be any suitable type of printer executing any suitable type of 3D printing process, or any other type of unit executing another additive manufacturing process. A variety of different types of additive manufacturing units 10102 and 3D printing processes are discussed below for purposes of illustration. However, the present disclosure is not limited to the 3D printing process described below.

In an embodiment, additive manufacturing unit 10102 may be configured to perform a Fused Deposition Modeling (FDM)™ process (e.g., also known as Fused Filament Fabrication™). The process of FDM may involve a software process capable of processing input files, such as stereolithography (STL) files. The object may be created, for example, by extruding small beads of thermoplastic material to form layers as the material hardens immediately after extrusion from a nozzle. Extrusion is a 3D printing technique in which a material, such as a polymer, metal (including alloy), etc., is propelled in fluid form through a tube into a moving nozzle, which extrudes the material to a target location, where the material subsequently cures in place. By moving the extruder continuously and precisely or starting and stopping it at very high speeds, the design is built layer by layer. The source material is typically supplied and stored in solid form, such as a filament or wire, which is wound into a coil and then heated with a heating element to bring the material into a thermoplastic state and flow the material between an "off" state and a maximum flow state. is extruded to supply material to an extrusion nozzle that can be controlled. A worm-drive, or any other suitable drive system, may be provided to propel the filament into the nozzle at a controlled rate. The nozzle is heated to melt the material. The thermoplastic material is heated past its state transition temperature (solid to fluid) and then deposited by an extrusion head. The nozzle can be moved both horizontally and vertically, for example by numerically controlled mechanisms. In embodiments, the nozzle may follow a tool path controlled by a computer-aided manufacturing (CAM) software package and the object is manufactured layer-by-layer, for example bottom-up.

In embodiments, the additive manufacturing unit 10102 may, for example, (a) perform rapid switching between source materials, such as facilitated by a set of valves, such as a set of high pressure valves, and/or (b) perform different operations on an item, such as comprising multiple source materials and multiple extrusion nozzles (and their support components, for example for movement and positioning), such as to enable simultaneous extrusion by multiple nozzles to enable simultaneous layering at a point. can do. In embodiments, additive manufacturing unit 10102 may perform voxelized soft materials printing and/or metal printing via multi-material, multi-nozzle printing, for example, with high-speed switching between materials at a rate of 50 or more times per second. makes possible.

In an embodiment, additive manufacturing unit 10102 may be configured to execute an electron beam freeform manufacturing (EBFFF) process. The EBFFF process can utilize electron beam welding technology to create metal parts. In embodiments, with the EBFFF method, metal preforms can be manufactured from computer-generated 3D drawings or models. Deposition paths and process parameters can be generated from post-processing of the virtual 3D model and executed by real-time computer control. Deposition takes place in a vacuum environment. The wire can be directed toward the melt pool and melted by a focused electron beam. The different parts of the object to be manufactured are built layer by layer by moving an electron beam and wire source across the surface of a base material, referred to as a substrate. The deposit solidifies immediately after the electron beam passes through it.

In an embodiment, additive manufacturing unit 10102 may be configured to perform a direct metal laser sintering process (DMLS). The DMLS process can involve a laser as a power source to sinter powdered materials, such as metals, at points in space defined by a 3D model, thus binding the materials together to create a solid structure. The DMLS process may involve the use of a 3D CAD model, from which a file, such as an .stl file, is created and transferred to software in additive manufacturing unit 10102. DMLS-based 3D printers can use high-power fiber lasers. Metal powders are fused into solid parts by locally melting them using a focused laser beam. Object parts are built layer by layer in a stratified manner.

In an embodiment, additive manufacturing unit 10102 may be configured to perform a selective laser melting (SLM) process. The SLM process uses 3D CAD data as a digital information source and uses energy in the form of a high-power laser beam to fuse fine metal powders together to create 3D metal parts. The process involves slicing the 3D CAD file data into layers to create a 2D image of each layer. A thin layer of atomized fine metal powder is uniformly distributed using a coating mechanism on a substrate plate clamped to an indexing table moving in the vertical (Z) axis. This occurs inside a chamber containing a tightly controlled atmosphere of an inert gas such as argon. Once each layer is dispensed, each 2D slice of the geometry is created by selectively applying laser energy to the powder surface by directing a focused laser beam using two high-frequency scanning mirrors in the X- and Y-axes. are fused. The laser energy allows complete melting of the particles to form solid metal. This process is repeated layer by layer until the part is complete. In embodiments, the SLM process may be a multi-scanner and/or multi-laser SLM process, such as enabling simultaneous action across multiple scans and/or multiple target points of a laser melting operation.

In embodiments, additive manufacturing unit 10102 may be configured to perform a selective thermal sintering process. The process may involve a thermal printhead applying heat to a layer of powder source material, causing it to become thermoplastic. Once a layer is complete, the powder bed of source material moves down, and automated rollers add a new layer of material, which is sintered to form the next cross-section of the object. Powder bed printing can refer to a technique in which one or more powders, typically metal powders, are joined through a variety of methods, such as laser or heat, to rapidly produce a final product. Typically, this is done by having areas filled with powder and connecting only the designed areas of powder while removing the rest layer by layer, or by adding powder layer by layer while simultaneously connecting the powders. Similar to photopolymerization, powder bed printing is significantly faster than other types of 3D printing. In embodiments, additive manufacturing unit 10102 may utilize multiple powder bed/roller subsystems, thereby creating a multi-material powder bed that allows simultaneous operation of different work target points and/or switching between materials. Makes application possible.

In embodiments, the various types of additive manufacturing units 10102 described herein may, for example, combine the advantageous properties (e.g., mechanical properties) of two materials to provide advantages over those of a single material. To do this, materials can be combined to produce an output that includes a composite of materials. In embodiments, the composite material produced in or by the additive manufacturing unit 10102 may include a functional gradient material (FGM), such as where two materials have a gradient interface that avoids a distinct boundary between the materials. is combined with This allows thermal and/or mechanical stresses arising from different material properties to be distributed over a larger volume/space, thereby mitigating problems such as cracking and failure that occur in non-gradient composite materials.

In embodiments, additive manufacturing unit 10102 may be configured to perform a selective laser sintering process. The selective laser sintering (SLS) process involves a laser that melts a flame-retardant plastic powder, which then solidifies to form the printing layer. In embodiments, additive manufacturing unit 10102 may be configured to perform a plaster-based 3D printing process. In embodiments, additive manufacturing unit 10102 may be configured to execute a laminate object manufacturing process. In this process, layers of adhesive-coated paper, plastic, or metal laminate can be continuously glued together and cut to shape with a knife or laser cutter. After the object has been manufactured by the additive manufacturing unit 10102, further modifications may be performed by machining or drilling after printing. In embodiments, selective laser sintering (SLS) involves multiple lasers, allowing switching and/or simultaneous operation of different target locations and/or different material types.

In an embodiment, additive manufacturing unit 10102 may be configured to perform a stereo-lithography (SLA) process. The process may use, for example, a resin from a vat of liquid ultraviolet curable photopolymer material, and an ultraviolet laser to build up the layers one at a time. For each layer, the laser beam traces a cross-section of the part pattern on the surface of the liquid resin. Exposure to ultraviolet laser light hardens and solidifies the pattern traced on the resin and bonds it to the underlying layer. In embodiments, the SLA process may involve multiple UV lasers, allowing switching and/or simultaneous operation on different target locations and/or different material types.

In an embodiment, additive manufacturing unit 10102 may be configured to implement digital light processing (DLP) methods. Digital light processing uses a projector to project an image of a cross-section of an object onto a vat of photopolymer (light-reactive plastic). Light selectively hardens only specific areas in the image. The printed layer is then repositioned to leave room for the uncured photopolymer to fill the newly created space between the print and the projector. Repeating this process builds the object one layer at a time. In embodiments, multiple DLP sources deliver light to different locations, allowing switching and/or simultaneous operation of different target locations within the light-reactive plastic material.

In embodiments, additive manufacturing unit 10102 may be configured to perform a photopolymerization method. In this process, droplets of liquid plastic are exposed to a laser beam of ultraviolet light. During this exposure, light converts the liquid into a solid. Photopolymerization is a technique in which a type of light is applied and/or a moving (e.g. laser) light source causes rising or falling layers of a photosensitive polymer to cure in areas that change over time as they rise or fall. Techniques are available to target different locations where the plastic material is located. This cures these areas of the polymer, and once the desired shape is created, the remaining uncured liquid polymer is removed, leaving the finished product. Photopolymerization is useful because of the speed with which the final product can be completed, with some types working up to 100 times faster than other 3D printing methods for some designs.

In embodiments, additive manufacturing unit 10102 may involve the use of an inkjet type printhead to deliver liquid or colloidal binder material to a layer of powdered building material. Printing techniques may involve applying a layer of powdered building material to a surface, for example using a roller. After the building material is applied to the surface, the printhead delivers liquid binder to predetermined areas of the layer of material. The binder penetrates the material and reacts with the powder, for example by activating the adhesive in the powder, causing the layer to solidify in the printed area. After the first cross-sectional portion is formed, the steps are repeated and successive cross-sectional portions are manufactured until the final product is formed.

In an embodiment, a method performed by additive manufacturing unit 10102 includes the deposition of successive layers of buildup material on a rotating buildup table and in a predetermined pattern on each successive layer of buildup material to form a 3D object. may involve deposition of a liquid.

In embodiments, additive manufacturing unit 10102 may include multiple types of additive manufacturing capabilities, among those described herein or understood by those skilled in the art, thereby forming a hybrid additive manufacturing unit. do. In embodiments, hybrid additive manufacturing units may further integrate other manufacturing capabilities such as cutting technologies, assembly systems, handling systems, finishing systems, etc. In embodiments, a hybrid additive manufacturing unit may integrate injection delivery of colloidal binder material with liquid polymerization technology.

In embodiments, platform 10110 may provide 3D printed products that conform to a user's body part/anatomy, including wearables such as eyewear, footwear, earwear, and headgear. In embodiments, suitability may be based on scans of body parts or anatomical features, such as laser or other structured light scans, MRI, EEG, computed tomography, ultrasound or other imaging scans, etc. The 3D topology of the anatomical features can be used as an input source for creation by a CAD system or other design system (which may be linked or integrated into the additive manufacturing platform) of the design for additive manufacturing. The design includes anatomically-compatible items that conform well to the anatomy (such as an audible unit that fits the inner ear, headgear that fits the head, braces that fit a joint, etc.) and/or items that are intended to replace a portion of the anatomy, such as a prosthesis. It can be configured to generate

In an embodiment, platform 10110 has the capability for self-starting and self-powering.

In embodiments, platform 10110 has built-in recycling capabilities where scrap portions can be automatically returned to the production process and support material and excess powder can be returned to the production process.

Figure 115 is a schematic diagram illustrating an example implementation of an autonomous additive manufacturing platform for automating and optimizing digital production workflows for additive manufacturing (e.g., metal manufacturing), according to some embodiments of the present disclosure.

Autonomous additive manufacturing platform 10110 includes a data collection and management system 10202, a data storage system 10204, and a data processing system 10206. The manufacturing workflow management application 10208 manages a variety of workflows, events, and applications related to printing and supply chain, including monitoring, inventory counting, queue management, storage management, production reporting, production analytics, and more.

Data collection and management system 10202 collects and organizes data collected from various data sources, including real-time data collected from a set of sensors. Some examples of sensors that provide data as input to the data collection and management system 10202 include power and energy sensors, mass sensors, position sensors, temperature sensors, humidity sensors, pressure sensors, viscosity sensors, flow sensors, chemical/ It includes gas sensors, strain gauges to be measured, image capture/cameras, video capture, thermal imaging, hyperspectral imaging, sound sensors and air quality sensors.

Data storage system 10204 may store a wide range of data types using a variety of storage media, data architectures, and formats, including, but not limited to, the following, among many other data types associated with platform 10110: (part profiles, entity or asset data (such as a product profile, printer profile), state data (such as indicating a state, condition state, or other indicator for any asset, entity, application, component or element of platform 10110), ( User data (including process events, financial events, transaction events, output events, input events, state changes) (including identity data, role data, task data, workflow data, health data, performance data, quality data and many other types) Events that occur within platform 10110, including operation events, workflow events, repair events, maintenance events, service events, damage events, replacement events, refueling events, recharge events, delivery events, supply chain events, etc. , or for any of a wide range of events, including operational data, transaction data, workflow data, maintenance data, and many other types of data, including events related to or related to one or more applications) data; Claims data (e.g., data relating to product liability, general liability, personal injury and other liability claims, and contracts such as supply contract performance claims, product delivery requirements, quality assurance claims, indemnity claims, delivery requirements, timing requirements, milestones, key performance indicators, etc.) billing data relating to); Accounting data (e.g. data on completion of contractual requirements, satisfaction of receivables, payment of duties and customs, etc.) and risk management (e.g. on parts or products supplied, quantities, prices, delivery, source, route, customs information, etc.) data.

In embodiments, data storage system 10204 may be used to maintain a serial or other record, over time, of an entity or asset, e.g., a part or product described herein or any other asset or entity. , data can be stored in distributed ledgers, digital threads, etc.

Data processing system 10206 includes an artificial intelligence system 10212, such as machine learning system 10210. Machine learning system 10210 may include analysis, simulation, decision making, and/or data processing, data analysis, simulation creation, and/or simulation analysis of one or more of the assets or entities of distributed manufacturing network 10130 of FIG. 114 . And a machine learning model 213 for performing predictive analysis can be defined. In embodiments, platform 10110 may include a set of artificial intelligence systems 10212 (including any of the types described in this application or in documents incorporated herein by reference), which include (a ) to operate on a set of inputs and/or a set of optimization factors to automatically select the appropriate type of additive manufacturing for the design/task; (b) to automatically discover a set of available additive manufacturing units 10102 (optionally including single-type units and/or hybrid type units); (c) to perform an additive manufacturing operation; to automatically select a set of; (d) automatically schedule a set of additive manufacturing units 10102 to perform a set of additive manufacturing operations; (e) automatically configure a selected set of additive manufacturing units 10102 to perform a set of additive manufacturing operations using a set of designs provided by a set of artificial intelligence systems; and/or (f) automatically organize the logistics and delivery of the set of outputs from the set of additive manufacturing units. In an embodiment, the set of inputs includes the location and type of available additive manufacturing units 10102, current work schedules for the additive manufacturing units, cost factors (e.g., material cost, energy cost, cost of IT resources, labor cost, pricing for manufacturing services, etc.), design inputs (e.g. functional requirements for strength, flexibility, elasticity, temperature tolerance, strain tolerance, wear resistance, water resistance, stress tolerance, weight bearing, tensile strength, load bearing, etc.), as well as as well as compatibility factors (including shape compatibility, biocompatibility, chemical compatibility, environmental compatibility, etc.). Optimization factors include aesthetic factors, compatibility factors (as mentioned above), economic factors (such as marginal cost, total cost, profitability, price, brand impact, etc.), various ongoing manufacturing, service, maintenance, marketing, etc. Timing factors (for coordination with workflows and activities, including delivery and/or logistics processes), prioritization factors, etc. In an embodiment, the artificial intelligence system of platform 10110 is trained based on a training data set that includes expert interactions with a set of additive manufacturing projects involving various types of additive manufacturing options. In embodiments, the AI system is trained based on outcome factors such as product quality and/or product defect outcomes, economic outcomes, timely completion outcomes, etc., including, for example, deep learning, supervised learning, and/or semi-supervised learning. In an embodiment, the AI system is distributed between the additive manufacturing unit 10102 and a host system, such as a cloud-based system. In an embodiment, the AI system is integrated into additive manufacturing unit 10102. In embodiments, the AI system is distributed across a set of additive manufacturing units 10102, such as a mesh or network of nodes of the additive manufacturing units 10102, and thus the above capabilities can be implemented, for example, (e.g., “Additive Manufacturing as a Service”) By self-organization of units 10102 in cooperation with other units, such as a fleet of additive manufacturing units 10102 owned and/or jointly operated by an enterprise and/or shared by a set of users (in a "manufacturing" system), Coordinated across units. As one of many possible examples, the AI system of platform 10110 takes a set of design requirements, such as functional requirements, generates a set of designs that satisfy the functional requirements, and stacks them to create each set of designs. Determine the optimal combination of manufacturing types, find and compare available additive manufacturing units for each combination (e.g., using economic and other factors), and select, configure, and schedule the units to perform the design. You can. For example, among the many possibilities across a wide range of product categories, AI systems can match colors from a large color palette to the customer's exact preferences; for example, latex can withstand impact and bending while also meeting design requirements for using biocompatible, water-resistant materials. -Functional requirements can be taken for customized wearable devices for individual users with allergies. The AI system automatically creates a set of instructions to create a wearable device using a combination/hybrid of photopolymerization (operating on non-latex polymers) for the components of the wearable that will be in contact with the user and DMLS processes for the internal metal/alloy components. It can be created with The AI system then finds available units, such as different units or integrated/hybrid units, schedules the units to perform tasks (e.g., to meet targeted delivery times), configures the units, and transfers the tasks. and schedule delivery. Accordingly, through the use of a set of additive manufacturing units, AI systems can design, create and deliver highly customized products based on customer-specific design requirements, including health requirements, physical construction requirements, economic factors, and preferences, among others. can be managed automatically.

In an embodiment, the AI system is implemented as an intelligence layer 140 that receives requests from a set of intelligence layer clients and responds to these requests by providing intelligent services (e.g., decisions, classifications, predictions, etc.) to these clients. .

In embodiments, machine learning model 10213 is an algorithm and/or statistical model that performs a particular task without using explicit instructions and instead relying on patterns and inferences. Machine learning model 10213 may build one or more mathematical models based on training data to make predictions and/or decisions without being explicitly programmed to perform a specific task. Machine learning model 10213 may receive input of sensor data or other data as training data, including event data and state data related to one or more of the entities or assets, or other inputs mentioned above or throughout this disclosure. You can. Sensor data input to machine learning model 10213 may be used for data processing, data analysis, simulation generation, and/or simulation analysis of one or more of the distributed manufacturing network entities or assets, simulations, decision making, and/or predictive analytics. It can be used to train a machine learning model 10213 to perform. Machine learning model 10213 may also use input data from a user or users of autonomous additive manufacturing platform 10110. In embodiments, machine learning model 10213 may use input data and sensor data to determine an optimal set of process parameters for 3D printing of a part by additive manufacturing unit 10102. Machine learning models (10213) include artificial neural networks, decision trees, logistic regression models, stochastic gradient descent models, fuzzy classifiers, support vector machines, Bayesian networks, hierarchical clustering algorithms, k-means algorithms, genetic algorithms, and any other suitable It may include a type of machine learning model, or a combination thereof. The machine learning model 10213 is configured to learn via supervised learning, unsupervised learning, reinforcement learning, self-learning, feature learning, sparse dictionary learning, anomaly detection, association rules, combinations thereof, or any other suitable algorithm for learning. It can be configured.

In embodiments, artificial intelligence system 10212 may define a digital twin system 10214 to create a digital replica, or digital twin, of one or more of the distributed manufacturing network entities. The digital twin of one or more of the distributed manufacturing network entities may use substantially real-time sensor data to provide a substantially real-time virtual representation of the distributed manufacturing network entity and a simulation of one or more possible future states of the one or more distributed manufacturing network entities. A digital twin exists simultaneously with one or more distributed manufacturing network entities (physical twins) from which it is replicated and contains sensor data, test and inspection results, maintenance performed, and modifications to reflect the current conditions or parameter values of one or more distributed manufacturing network entities. It can be continuously updated based on etc. A digital twin, in embodiments, provides one or more simulations of both the physical elements and characteristics of one or more distributed manufacturing network entities being replicated and their dynamics throughout the lifecycle of the one or more distributed manufacturing network entities being replicated. The digital twin can be used to identify periods when component wear can be a problem, for example, during high stresses, during the design phase before one or more entities are manufactured or manufactured, or during or after the construction or manufacture of one or more entities. To simulate the state of one or more distributed manufacturing network entities after a period of time, during peak throughput operation, after one or more hypothetical or planned improvements have been made to the one or more distributed manufacturing network entities, or during any other suitable hypothetical situation. By allowing virtual extrapolation of sensor data, it can provide a virtual simulation of one or more distributed manufacturing network entities. In an embodiment, machine learning model 10213 can, for example, predict possible improvements to one or more distributed manufacturing network entities, predict when one or more components of one or more distributed manufacturing network entities may fail, and /or by suggesting to one or more distributed manufacturing network entities possible improvements, such as changes to parameters, arrangements, components, or any other suitable changes to the distributed manufacturing network entities, thereby creating a virtual situation for simulation using the digital twin. It can be predicted automatically.

A digital twin allows simulation of one or more distributed manufacturing network entities during both the design and operation phases of the one or more distributed manufacturing network entities, as well as the simulation of virtual operating conditions and configurations of one or more distributed manufacturing network entities. A digital twin is a digital twin of one or more distributed manufacturing network entities, including temperature, pressure, wear, light, humidity, strain, expansion, contraction, deflection, bending, stress, deformation, load-bearing, and shrinkage within, on, and around each of the entities. It allows analysis and simulation of one or more distributed manufacturing network entities by facilitating the observation and measurement of almost any type of metric. Insights gained from analysis and simulation using digital twins can be passed on to design or manufacturing processes to improve these processes.

In embodiments, machine learning model 10213 may process sensor data, including event data and state data, to define simulation data for use by digital twin system 10214. Machine learning model 10213 may, for example, receive state data and event data related to a particular distributed manufacturing network entity, and store the state data and event data in the creation of a digital replica of the distributed manufacturing network entity. A series of operations can be performed on state data and event data to format it into a format suitable for use. For example, one or more distributed manufacturing network entities may include products manufactured by additive manufacturing unit 10102. The machine learning model may collect data from one or more sensors located on, near, within, and around the product. The machine learning model may perform operations on the sensor data to process the sensor data into simulation data and output the simulation data to the digital twin system 10214. The digital twin system 10214 may use the simulation data to create one or more product twins 10215, where the simulation measures metrics including, for example, temperature, wear, speed, rotation, and vibration of the product and its components. Includes. The simulation may be a substantially real-time simulation that allows users of platform 10110 to view a simulation of the product, metrics related thereto, and metrics related thereto in substantially real time. A simulation may be a predictive or hypothetical situation that allows a user of platform 10110 to view a predictive or hypothetical simulation of a product, metrics associated therewith, and metrics associated with its components.

In embodiments, machine learning model 10213 and digital twin system 10214 process sensor data and create a digital twin of a set of distributed manufacturing network entities to enable design, real-time simulation, and predictive simulation of related groups of distributed manufacturing network entities. , and/or may facilitate virtual simulation.

In embodiments, control system 10216 within data processing system 10206 may adjust process parameters of a 3D printing process in real time based on simulation.

In embodiments, a distributed manufacturing network entity, such as an additive manufacturing unit 10102 or a platform 10110, may optionally automatically create a set of digital twins of a set of manufactured items, such as products, components, parts, etc. In embodiments, a digital twin of a manufactured item created by an additive manufacturing unit 10102 or platform 10110 may include, be linked to, enriched by, and/or integrated with, among other things: : (a) a set of instructions for additively manufacturing the item, such as including shape information, material layering information, functional information, operating parameter information (as described elsewhere in this application), etc.; (b) a training data set on which an artificial intelligence system is trained in relation to the design or manufacture of an item; (c) accurate manufacturing of the item, e.g., in the case of an item, linking a series of images of the layers of the item as it was created with data indicating the environment in which the item was manufactured, the equipment or tools used, the materials used, etc. Time series sensor data (e.g., imaging data from various imaging systems) indicative of conditions; Temperature, pressure, fluid flow rate, heat flux data, volume data, topology data, radiation data (e.g. intensity of laser, visible light, infrared light, UV, x-ray, magnetic field, electric field, etc.), sensor data sets, such as those containing chemical information (e.g., presence of reactants, catalysts, etc.), biological data (e.g., presence and status of biomaterials, pathogens, and other agents), etc.; (d) results of testing before, during or after manufacturing, such as equipment testing, material testing, stress testing, visual inspection (including by machine vision), strain testing, torsion testing, load testing, impact testing, motion testing, etc. A test data set such as that representing; (e) manufacturing information regarding similar items, such as manufacturing results, usage, etc.; etc. In embodiments, additive manufacturing unit 10102 may automatically create a digital twin upon receiving instructions to manufacture an item and subsequently enrich and/or modify the digital twin during and/or after manufacturing. In an embodiment, additive manufacturing unit 10102 may transfer the above-referenced data about the digital twin of the item to the item (e.g., by writing to a data structure embedded in or placed on the item, such as a chip). It can be automatically embedded on tags, containers, or packages.

116 is a block diagram illustrating information flow in an autonomous additive manufacturing platform 10110 for optimization of different operating parameters of an additive manufacturing process, according to some embodiments of the present disclosure. In embodiments, parameters may be associated with a 3D printed part, 3D printed product, 3D printing process, or 3D printing machine. Some examples of parameters include: extrusion temperature, material deposition rate, tool path, voltage setting of the heating device, exposure pattern, layer height, printing surface temperature, layer height/thickness, build speed, build material flow rate, part. Contains orientation, geometry and volume information about air gaps, pores, voids, pores, lumens, clearances, conduits, etc., support structure settings, ambient conditions including temperature, humidity and pressure, raw material conditions including temperature and viscosity, temperature Part conditions, stress concentrations including compressive, tensile, shear, bending and torsional stresses. Again, the parameters are typically specific to a given additive manufacturing technology, material, geometry and application, or a particular hybrid or combination thereof.

Referring to Figure 116, at 10300, input data for printing a product is received at autonomous additive manufacturing platform 10110. Input data may be received in the user interface of platform 10110 and may include details such as 3D printing technology, product geometry and key features, printing materials, etc. In embodiments, the input data may only include the desired properties of the product (such as strength, stiffness, yield, elasticity, elongation, electrical conductivity, thermal conductivity, etc.) or application area (aerospace, dental, automotive, jewelry, etc.). and platform 10110 can determine details such as 3D printing technology or materials to be used for printing. This may be done automatically (e.g., by artificial intelligence), or with human interaction and/or supervision, such as when a set of recommended details is proposed by AI and confirmed and/or modified by a human user. .

At 10302, an instruction set for additive manufacturing, such as a 3D printing profile, is determined based on inputs received at 10300 as well as simulations received from machine learning system 10210 and digital twin system 10214. The profile contains parameters for additive manufacturing of products using, for example, a 3D printer.

At 10304, sensor data (including but not limited to ambient, product or material temperature); compressive, shear, tensile, bending and torsional stresses; oxygen, carbon dioxide levels, and ozone levels; humidity; vibration; sound signatures and visual indicators) are collected from additive manufacturing (e.g., 3D printing) processes. Data collection and management system 10202 helps collect sensor data through an array of sensors and other data collection technologies, such as IoT devices, machine vision systems, etc. Collected data may be analyzed at an edge device or transmitted to one or more data pools within data storage system 10204 for later consumption, for example, by local or remote intelligence. Within the computing infrastructure proximate to the additive manufacturing unit(s) 10102 (e.g., in a local area network of a building, campus, or other premises where the additive manufacturing unit(s) 10102 is located, and/or to the additive manufacturing unit(s) 10102 (in a connected vehicle transporting 10102) and/or integrated with or included in additive manufacturing unit 10102, such as when additive manufacturing unit 10102 has onboard edge computing and/or connectivity resources, such as 5G (or other The use of cloud-connectable edge devices, such as those with cellular, Wifi, Bluetooth, fixed networking resources, etc., benefits from the extensive computing and data storage capabilities provided by highly scalable cloud computing resources, such as servers. provides the opportunity to provide rapid, real-time or near-real-time processing responsiveness.

In embodiments, data may also be stored in a blockchain such that storage is distributed across multiple manufacturing nodes as well as other data storage devices or systems. In embodiments, this may take the form of a distributed ledger that can capture transactions, events, etc., such as financial events involving additive manufacturing, smart contract related events, operational events (such as scheduling or completion of a task), etc. Data may also be multiplexed using sensor fusion or otherwise condensed, relayed over a network, and fed to a machine learning system that utilizes one or more machine learning models.

At 10306, parameters may be dynamically adjusted as needed based on analysis of sensor data. As 3D printing is completed, data related to the results of the 3D printing process is collected at 10308. Resulting data may be collected through a user interface that provides the user with information regarding the success or failure of the 3D print. The data is then provided as feedback to a machine learning system 10210 that uses the feedback to train or improve an initial machine learning model (such as improving by adjusting weights, rules, parameters, etc. based on the feedback). . In embodiments, the feedback is utilized to analyze trends for multiple 3D printing operations performed by one or more users across multiple additive manufacturing units 10102 and manufacturing nodes 10100.

In an embodiment, autonomous additive manufacturing platform 10110 uses machine learning to provide optimization and process control across the entire lifecycle of manufacturing, from product concept and design through manufacturing and distribution to service and maintenance.

In an embodiment, autonomous additive manufacturing platform 10110 provides generative design and topology optimization to determine at least one product design suitable for manufacturing.

In embodiments, autonomous additive manufacturing platform 10110 provides optimization of the construction preparation process.

In embodiments, autonomous additive manufacturing platform 10110 optimizes the part orientation process for superior production results.

In embodiments, autonomous additive manufacturing platform 10110 automatically determines and recommends support structures to minimize material cost, printing time, post-processing, and risk of damage to 3D printed parts (when supports are removed). .

In an embodiment, autonomous additive manufacturing platform 10110 provides for optimizing tool path generation. For example, in a 3D printer, the tool path may include the trajectory of the nozzle and/or print head. In embodiments, tool path generation allows the manufacturing process to fill the boundaries and interior regions of each sliced layer. Different types of tool path strategies and algorithms are possible, such as zigzag, contour, spiral, and partition patterns, taking into account the construction time, cost, geometric quality, distortion, shrinkage, strength, and stiffness of the manufacturing model. In embodiments, an artificial intelligence system may be trained on results such as those described above to provide recommended tool paths and/or fully automate tool path creation.

In embodiments, autonomous additive manufacturing platform 10110 provides optimized dynamic 2D, 2.5D, and 3D nesting to maximize the number of printed parts while minimizing raw material waste. In embodiments, nesting algorithms evaluate individual part priorities to ensure that high priority parts are treated accordingly, e.g., scheduling priority, quality priority, usability priority, positioning priority, etc. This is optimized. In embodiments, nesting is optimized such that the nesting algorithm minimizes the travel time of the cutting tool. In embodiments, nesting is optimized such that the nesting algorithm is integrated with support structure optimization.

In embodiments, autonomous additive manufacturing platform 10110 provides optimization of post-processing processes.

In an embodiment, autonomous additive manufacturing platform 10110 provides an automated powder removal system that utilizes a digital twin, where the digital twin calculates optimal movement of the powder removal system during powder removal.

In embodiments, autonomous additive manufacturing platform 10110 provides automated, hands-free support structure removal.

In an embodiment, autonomous additive manufacturing platform 10110 provides automated surface finishing.

In embodiments, autonomous additive manufacturing platform 10110 provides automated part metrology for use with integrated quality and process control systems.

In embodiments, the manufacturing methods described herein may use material additives during processing that impart various properties to the finished part. Examples in plastic injection molding include glass fibers for added strength, and electrically conductive and shielding fibers for tailored electrical properties. In some applications, the orientation of added fibers or other materials can affect the performance of the finished part. For example, in glass fiber reinforced applications, long fiber orientation can dictate the minimum and maximum strain orientation under stress. Fiber orientation during manufacturing can only be partially controlled through mold design, injection nozzle position and pressure, and other process controls.

3D printed parts can also be manufactured using material additives; However, most 3D printing methods can only produce materials with limited ability to optimize layer properties, such as fiber orientation, to help optimize finished part performance. For example, 3D printers can use nozzles to extrude a variety of plastic materials, but the inherent flow characteristics of fixed nozzles, and the limitations of the 3D printing process in general, limit options for materials engineering the finished part. The use of these 3D printing nozzles provides the ability to control the orientation of the layered material as it is placed for part production. These developments provide opportunities to fine-tune material performance, for example, localized orientation for structural enhancement, or homogeneous random orientation for electrical shielding performance. In an example, this capability may be provided by a 3D printing nozzle that uses actuated flexible elements to change the shape of the nozzle during material application, resulting in predictable fiber orientation. This can be used in conjunction with other printing process parameters such as nozzle orientation, flow rate and pressure to further improve material properties. Example use cases include, but are not limited to: one or more engineering properties that can be varied across a single part to provide targeted performance, e.g., varying stiffness; Optimized use of materials based on improved process control, e.g. using less material to produce parts with the same functional performance, and the combined ability to combine with random conductive additives for specified electrical performance; For example, providing control of multiple additives to impart orientation of structural long fibers for structural performance.

Artificial intelligence, machine learning, automation (including robotic process automation, remote control, autonomous behavior, automated configuration, etc.), expert systems, self-organization, and adaptive intelligent systems for prediction, classification, optimization, etc. Embodiments of the present disclosure, including those that benefit from the use of neural networks, such as trained neural networks, for pattern recognition, classification of one or more parameters, characteristics, or phenomena, support of autonomic control, and for other purposes. You can get it.

Neural networks (or artificial neural networks) are a family of statistical learning models that derive their inspiration from biological neural networks, can rely on large numbers of inputs, and are commonly used to estimate or approximate unknown functions. A neural network represents a system of interconnected “neurons” that send messages to each other. Connections have numerical weights that can be tuned based on experience, allowing the neural network to adapt to input and learn.

Throughout this disclosure, references to artificial intelligence, neural networks, or neural networks refer to a wide range of different types of machine learning systems, neural networks such as feedforward neural networks, convolutional neural networks (CNNs), recurrent neural networks (RNNs), long short-term memory (LSTM) neural networks; Gated recurrent unit (GRU) neural network, self-organizing map (SOM) neural network (e.g., Kohonen self-organizing neural network), autoencoder (AE) neural network, encoder-decoder neural network, modular neural network, or a variant, tactic. Hybrids or combinations of one, or rule-based systems, such as reinforcement learning (RL) systems or other expert systems, and model-based systems (including those based on physical models, statistical models, flow-based models, biological models, biomimetic models, etc. It should be understood to include combinations with).

The neural network described above may have various nodes or neurons, which may perform various functions on input, such as input received from a sensor or other data source, including other nodes for predicting one or more outputs. A function may involve weights, features, feature vectors, etc. Neurons may include perceptrons, neurons that mimic biological functions (such as the human senses of touch, sight, taste, hearing, and smell). Neural networks may utilize multiple operational layers, including one or more hidden layers located between the input layer and the output layer. The output of each layer can be used as input to another layer, for example the next hidden layer or output layer. The output of a particular neuron may be a weighted sum of the inputs to the neuron, adjusted by the bias and multiplied by an activation function, such as a rectified linear unit (ReLU) or sigmoid function.

In many embodiments, an expert system or neural network may be trained, such as by a human operator or supervisor, or based on a data set, model, etc. Training a neural network involves providing input to an untrained neural network to produce a predicted output, comparing the predicted output to the expected output, and adjusting the algorithm to take into account the differences between the predicted and expected outputs. This may involve updating weights and biases. Specifically, a cost function can be used to calculate the difference between predicted and expected output. By calculating the derivative of the cost function with respect to the network's weights and biases, the weights and biases can be adjusted iteratively over multiple cycles to minimize the cost function. Training may be completed when the predicted output satisfies a convergence condition, e.g., a calculated cost of small size, as determined by a cost function.

Training involves not only one or more training data sets representing values (including many of the types described throughout this disclosure), but also one or more sets of results, such as the result of a process, the result of a computation, the result of an event, the result of an activity, etc. It may involve presenting indicators to a neural network. Training a neural network to optimize one or more systems based on one or more optimization approaches such as Bayesian approach, parametric Bayes classifier approach, k-nearest-neighbor classifier approach, iterative approach, interpolation approach, Pareto optimization approach, algorithmic approach, etc. This may include training in optimization, such as training . Feedback may be provided in a process of variation and selection, such as using a genetic algorithm to evolve one or more solutions based on feedback over a series of rounds.

In embodiments, receiving data streams and other inputs collected in one or more environments (e.g., by a mobile data collector) and transmitted to a cloud platform over one or more networks, including using network coding to provide efficient transmission. Multiple neural networks may be deployed on a cloud platform. In the cloud platform, a plurality of different neural networks of various types (including modular form, structurally adaptive form, hybrid, etc.), optionally using massively parallel computing power, perform prediction, classification, and control functions, and the present disclosure It may be used to provide other outputs as described with respect to the expert system disclosed throughout. Different neural networks can be selected, for example by an expert system, for a particular task involved in a given situation, workflow, environmental process, system, etc., so that an appropriate type of neural network with appropriate input set, weights, node types and functions, etc. They can be structured to compete with each other (optionally including using evolutionary algorithms, genetic algorithms, etc.).

In embodiments, the methods and systems described in this application involving expert systems or self-organizing capabilities may use feedforward neural networks, which generate a series of neurons or nodes from data input, such as a source of data about an individual. Through this, information is moved in one direction to the output. Data can move from input nodes to output nodes, and optionally through one or more hidden nodes without looping. In embodiments, a feedforward neural network may be composed of various types of units, such as binary McCulloch-Pitts neurons, the simplest of which is a perceptron.

In embodiments, artificial intelligence and machine learning systems in the data processing system of autonomous additive manufacturing platform 10110 may enable automatic classification and clustering of 3D printed parts and products. In embodiments, artificial intelligence and machine learning systems in the data processing system of autonomous additive manufacturing platform 10110 may enable automatic classification and clustering of malicious defects in the additive manufacturing process.

The automated parts and defect classification methods and systems of the present disclosure may be implemented using image sensors and/or machine vision systems. Machine vision systems can monitor the additive manufacturing process in real time, for example, by capturing and analyzing images of the part or other item being printed. Automated image processing of the captured images can then be used to determine various part characteristics, such as dimensions (overall dimensions or dimensions of specific features), feature angles, feature areas, and surface finish (e.g., degree of light reflectivity, per unit area). number of pits and/or scratches), etc. Machine vision systems also track the process to detect any defects or errors in the printed part in real time while successive layers of material are being deposited by the 3D printer.

Defects are identified, for example, by removing noise from inspection data and subtracting a reference data set (e.g., a reference image of defect-free areas if a machine vision tool is being used for inspection), and individual objects. Use unsupervised machine learning algorithms, such as cluster analysis or artificial neural networks, to classify defects as meeting or not meeting a specified set of decision criteria (e.g., decision boundaries) in the feature space being monitored. It can be classified as follows. For example, a partially printed part can be compared to a rendering of the partial part, and if the partial part differs from the rendering by more than a selected threshold, the part can be classified as defective.

In embodiments, in-process defect classification data may be a set or sequence of process control parameter adjustments to implement corrective action, for example, to adjust layer dimensions or thickness to correct a defect when first detected. Can be used by machine learning algorithms to determine . In some embodiments, in-process automated defect classification may be used by machine learning algorithms to send warning or error signals to operators, or, optionally, to automatically stop the deposition process.

In an embodiment, a machine vision system uses a variable focus liquid lens-based camera for image capture and defect detection. In embodiments, the machine vision system uses infrared or visible wavelength cameras.

In an embodiment, the data processing system is implemented as an intelligence layer 140 that uses neural networks to provide real-time adaptive control of the additive manufacturing process, including part defect classification and feedback.

In some embodiments, a neural network model may be used directly to determine adjustments to handle control parameters using training or learning of the neural network model. Initially, the model is allowed to randomly select from a range of values for each input process control parameter or action. If a process control parameter adjustment or sequence of actions leads to a defect or defect, it is scored as leading to an undesirable (or negative) result. Repetition of the process using a different set of randomly selected values for each process control parameter or action leads to consolidation of the sequence resulting in a desired (or positive) result. Ultimately, the neural network model “learns” what adjustments to make to the set or sequence of deposition process control parameters or actions to achieve the target result, i.e., a defect-free printed part.

In embodiments, the methods and systems described in this application may use convolutional neural networks (referred to in some cases as CNNs, ConvNets, shift invariant neural networks, or spatially invariant neural networks), the units of which are similar to the visual cortex of the human brain. Connected by patterns. For example, CNNs can provide automatic classification and clustering of parts and defects in additive manufacturing processes.

In embodiments, one or more models built on the basic framework of convolutional neural networks may be used. For example, an object detection model can be used that extends the functionality of CNN-based image classification models by not only classifying parts or defects but also determining their location in the image in terms of bounding boxes. Similarly, Region-based CNN (R-CNN) models can be used to extract regions of interest (ROIs), where each ROI is a rectangle that can represent the boundary of a portion within the image.

In embodiments, a capsule network may be used to achieve similar classification performance of a CNN using fewer labeled training examples.

In embodiments, a transformer-based encoder-decoder architecture using an attention mechanism may be used in conjunction with or instead of a convolutional neural network.

117 is a schematic diagram illustrating a system for learning based on data from platform 10110 to train an artificial learning system to use a digital twin for classification, prediction, and decision-making in accordance with some embodiments of the present disclosure. .

117, the digital twin system 10214 within the autonomous additive manufacturing platform 10110 includes a product twin 10215, a part twin 10504, and a printer twin (10215) that allow for modeling, simulation, prediction, decision-making, and classification. 10506), user twin (10508), manufacturing node twin (10510), packager twin (10512), etc. The digital twin system 10214 may be populated with related data, for example, the product twin 10215 may be populated with data related to the corresponding product, including dimensional data, material data, feature data, thermal data, pricing data, etc. You can.

In embodiments, a digital twin may be created from another digital twin. For example, product twin 10215 may be created using one or more part twins 10504. In another example, part twin 10504 may be created using product twin 10215. In embodiments, a digital twin may be embedded in another digital twin. For example, partial digital twin 10504 may be embedded in a product digital twin 10215, which may be embedded in a manufacturing node digital twin 10510.

In embodiments, simulation management system 10514 may set up, provision, configure, and otherwise manage simulations and interactions between digital twins 10214.

In embodiments, artificial intelligence system 10212 is configured to run simulations in simulation management system 10514 using part twin 10502 and/or other digital twins available to digital twin system 10214. For example, artificial intelligence system 10212 may adjust one or more characteristics of printer twin 10506 as a set of part twins 10504 are printed by a 3D printer. In an embodiment, artificial intelligence system 10212 may, for each set of features, run a simulation based on the set of features and collect simulation result data resulting from the simulation. For example, when running a simulation on a set of part twins 10504 being manufactured on printer twin 10506, artificial intelligence system 10212 may vary the characteristics of printer twin 10506 and simulate the results to generate You can run . During the simulation, artificial intelligence system 10212 may vary the ambient temperature, pressure, humidity, lighting, and/or any other characteristics of printer twin 10506. In this example, the result may be the condition of component twin 10504 after high temperature has been applied. Results from the simulation can be used to train a machine learning model 10213. In embodiments, machine learning system 10210 may receive training data, results data, simulation data, and/or any other data from other data sources 10114. In embodiments, machine learning system 10210 may train/strengthen machine learning model 10213 using the received data to improve the model.

In embodiments, machine learning system 10210 may be used to make product classifications, predictions, recommendations, and/or decisions or instructions governing design, construction, material selection, shape selection, manufacturing type, job scheduling, etc. and train one or more models that are used by artificial intelligence system 10212 to generate or facilitate decisions or instructions related to the part.

In an example embodiment, artificial intelligence system 10212 trains a partial failure prediction model. A failure prediction model may be a model that receives part-related data and outputs one or more predictions or answers regarding the probability of part failure. Training data can be collected from multiple sources, including part specifications, environmental data, sensor data, machine vision data, and outcome data. Some examples of questions that predictive models can answer are: when will the machine fail, what type of failure will it be, what is the probability that the failure will occur in the next . Artificial intelligence system 10212 may train one or more predictive models to answer different questions. For example, a classification model can be trained to predict failure within a given time window, while a regression model can be trained to predict the remaining useful life of a machine. In embodiments, training may be done based on feedback received by the system, also referred to as “reinforcement learning.” Artificial intelligence system 10212 may receive the set of circumstances that led to the prediction (e.g., properties of the part, properties of the model, etc.) and results related to the part and may update the model according to the feedback.

In embodiments, artificial intelligence system 10212 may use clustering algorithms to identify failure patterns hidden in failure data to train models for detecting non-characteristic or abnormal behavior. Failure data and its historical history across multiple components can be clustered to understand how different patterns correlate with specific wear behavior. For example, if a failure occurs early in printing, the failure may be due to an uneven printing surface. If a failure occurs late in printing, the part may detach from the printing surface and the failure may be due to poor bed adhesion and/or distortion. All information collected can be used as feedback to the model. Over time, different failure modes will be associated with corresponding parameters. For example, poor bed adhesion is likely caused by incorrect temperature settings or printing orientation. Any failure to meet dimensional tolerances is likely caused by incorrect acceleration, velocity, or floor height. Machine learning system 10212 may determine the degree of correlation between each input and each failure mode.

In embodiments, artificial intelligence system 10212 may be used to initiate maintenance or replacement as needed, including platform-wide maintenance management, and as part of a computerized maintenance management system (MMS) for cutting tools. , filters and mechanical lasers can be configured to monitor them. In embodiments, additive manufacturing entities in a value chain network may be prepared, configured, and/or deployed to support replacement of components. a mobile additive manufacturing unit and/or a unit located in sufficient proximity to the service visit to facilitate rapid delivery of items produced by the additive manufacturing unit, for example, in connection with a service visit to a home or business; The same additive manufacturing unit may be designated to support service visits. Based on the nature of the service visit (e.g., type of equipment to be serviced, nature of component parts and materials within the equipment, problems identified, etc.), the additive manufacturing unit may select a variety of possible replacement parts, specializations, etc. to support the service visit. The tool may be equipped with suitable materials, such as a combination of metal printing materials and other printing materials suitable for printing other elements. In embodiments, the platform may include input from or related to a service visit, such as input indicative of the item being serviced (e.g., technical specifications, CAD designs, etc.); An input indicating a diagnosed problem (e.g., need to replace entire sub-assembly, need to repair cracks or other damage, etc.); It may take input captured by cameras, microphones, data collectors, sensors, and other information sources associated with a visit to the Service. For example, a service technician can capture a set of photos showing damaged areas. In embodiments, the platform may process input using an artificial intelligence system (such as a robotic process automation system trained on a training set of professional service visit data) to determine recommended actions, such as may involve replacement of parts and/or repair of parts. In some such embodiments, the platform determines whether replacement parts are readily available (e.g., using artificial intelligence systems such as robotic process automation trained on expert data sets) and/or whether the additive manufacturing system is It can automatically determine whether replacement parts need to be produced, to reduce costs, etc. Similarly, the platform may, in some embodiments, allow elements to be additively manufactured to facilitate repair, for example, where complementary components can be created using similar systems to replace worn or missing elements. You can automatically decide what should be done. In embodiments, an automated decision may be made by capturing a set of photographic images from a service visit, comparing them to a reference design for the applicable part, and adhering the part to the defective element (e.g., with a specified adhesive) to render the part compliant with the reference design. This can be accomplished using a machine vision system that generates a set of instructions for additively manufacturing complementary elements that can be added. In any such embodiment of recommending or configuring instructions for additive manufacturing, the platform discovers available units, configures instructions, initiates additive manufacturing, and provides updates, such as updates regarding when an element will be ready for use. can be provided to the service technician. In embodiments, the platform may, e.g., through trained AI agents, monitor the state of services and other related entities involved in other workflows, e.g., the overall planned duration of the service operation (e.g., outputs that will not be used immediately). (to allow de-prioritization of additive manufacturing jobs to be created), what other work is being done (e.g. to allow proper sequencing of additive manufacturing outputs that align with the overall workflow), and prioritization of service work (e.g. Automatically configure sets of tasks across a set of additive manufacturing units while recognizing, for example, whether they involve mission-critical items of operating equipment versus non-critical accessory items, the cost of downtime, or other factors. and can be scheduled. In embodiments, optimization of workflow across a set of additive manufacturing entities may be achieved by having an artificial intelligence system perform a set of simulations, such as simulations involving alternative scheduling sequences, design configurations, alternative output types, etc. In embodiments, simulations may be performed on additive manufacturing and other manufacturing entities (e.g., those performing cutting, drilling, etc.), including handoffs between sets of different manufacturing entity types, such as when handoffs are handled by a robotic handling system. may include sequences involving cutting manufacturing entities and/or finishing entities that perform finishing, hardening, etc. In an embodiment, a set of digital twins may be comprised of various manufacturing systems, various handling systems (robotic systems, arms, conveyors, etc., but also the human workforce), and/or the surrounding environment (e.g., vehicles, manufacturing facilities, campuses, or even cities). can represent the properties and capabilities of larger entities such as

In embodiments, artificial intelligence system 10212 may be configured to manage real-time dynamics affecting inventory levels for smart inventory and material management. This may include, for example, predicting inventory levels based on the various types of demand factors and/or sets of supply factors described herein and producing items for locations where shortages are expected. ) may include configuring a schedule for.

In embodiments, artificial intelligence system 10212 may be configured to build, maintain, and provide libraries of parts with pre-configured parameters that may be searchable by material, property, part type, part class, industry, compliance, etc. . This includes, for example, a set of search algorithms that discover parts by referencing published materials, including website materials, product specifications, etc.; A set of algorithms for querying a part provider's API or other interface, such as to query a database for part information; and/or a set of data acquisition systems that capture images, sensor data, test data, etc. of or about the part.

In embodiments, artificial intelligence system 10212 may be configured to analyze usage patterns associated with one or more users and learn user preferences with respect to output, timing, material, color, shape, orientation, and/or printing strategy. . For example, system 10212 may determine what materials were used for manufacturing, what processes were used for manufacturing, e.g., by additive manufacturing unit 10102, by location, by user, by organization, by role, etc. A profile can be developed that shows what was used, what shapes were produced, what finishing steps were performed, what colors were used, what functions were enabled, etc. Profiles can be used to determine, infer, or suggest preferences of users, organizations, etc. For example, an organization's preferred brand colors can be recognized, and matching materials and coatings are therefore recommended and/or pre-configured in the development of an additive manufacturing step.

In embodiments, artificial intelligence system 10212 may be configured to perform real-time calibration for one or more 3D printers. This may include training on a training dataset of expert users' calibration interactions. Calibration may, for example, cause artificial intelligence system 10212 to calibrate additive manufacturing unit 10102 to operate with a particular material, which may include material from a particular bin or lot of the same general type of material. By training, it can be job-specific.

In embodiments, artificial intelligence system 10212 may be configured to minimize material waste generation during an additive manufacturing process. These include structuring production to minimize material that needs to be removed in finishing stages, structuring production to produce output where unused material is easily removed for reuse, and/or producing reusable/recyclable materials. This may include structuring production to favor

In embodiments, artificial intelligence system 10212 may be configured to detect cybersecurity risks and threats to platform 10110.

In embodiments, artificial intelligence system 10212 may be configured to evaluate regulatory compliance. For example, in an embodiment, artificial intelligence system 10212 searches a library or other source of approved or certified product designs, such as those that are UL or CE recognized, FDA approved, OSHA approved, etc., and determines whether the output of the additive manufacturing is compliant. /Can be configured to compare design configurations to these to ensure that they will yield a product in the approved form. In embodiments, the artificial intelligence system 10212 may operate with a digital twin system, simulation system, etc. to simulate the performance of the resulting output, and determine the simulated performance to withstand forces, chemical effects, biological effects, radiation, etc. Comparison may be made with regulatory or other requirements such as those applicable to competency. For example, if a product component, such as a housing, is intended to provide shielding from radiation, artificial intelligence system 10212 may determine whether the product material, thickness, and geometry provide sufficient shielding to meet regulatory and/or design requirements. It can operate on or within a digital twin containing a radiation propagation physics model to automatically evaluate whether

In embodiments, artificial intelligence system 10212 may be configured to optimize power consumption for platform 10110. This involves (a) measuring the power consumed by various available activities; (b) training an artificial intelligence system 10212 to perform scheduling of additive manufacturing operations according to a predictive model of energy prices; and/or (c) causing the artificial intelligence system to perform large-scale body simulations to select desirable sequences of operations that produce advantageous power consumption patterns. It may include a training step.

In embodiments, models trained by machine learning system 10210 may be used by artificial intelligence system 10212 to run simulations on part twins to predict part shrinkage or expansion. This allows the artificial intelligence system 10212 to create a physical model that includes coefficients of thermal expansion for elements, alloys, compounds, mixtures, and/or combinations, in embodiments, comprising graded layers of materials without clear boundaries between the materials. This may include making use of a set. In embodiments, artificial intelligence system 10212 may be trained based on observed contraction and/or expansion during manufacturing and/or use.

In embodiments, models trained by machine learning system 10210 may be used by artificial intelligence system 10212 to run simulations on part twins to predict part warpage. This allows the artificial intelligence system 10212 to create a physical model that includes coefficients of thermal expansion for elements, alloys, compounds, mixtures, and/or combinations, in embodiments, comprising graded layers of materials without clear boundaries between the materials. This may include making use of a set. In embodiments, artificial intelligence system 10212 may be trained based on observed distortion during manufacturing and/or use.

In embodiments, the model trained by machine learning system 10210 uses artificial intelligence to run simulations on part twins to calculate necessary changes to the 3D printed process to compensate for part shrinkage, expansion, and/or warping. May be utilized by system 10212.

In embodiments, models trained by machine learning system 10210 may be used by artificial intelligence system 10212 to run simulations on part twins to test compatibility of additively manufactured parts. In embodiments, compatibility may be tested with one or more other parts within the assembly. In embodiments, compatibility may be tested with the operating environment. In embodiments, compatibility may be tested with a 3D printer. Compatibility refers to shape compatibility (e.g., key-in-lock; housing-around-interior; peg-in-hole; male-with-female, support-with-supported, or other type of interface/interconnect compatibility); Environmental compatibility (e.g., compatibility of the material with the expected use environment such as chemical factors, physical factors, radiation factors, biological factors, temperature, pressure, etc.); Functional compatibility (e.g., ability to withstand load, stress, torsion, etc.) may be included.

In embodiments, models trained by machine learning system 10210 may be used by artificial intelligence system 10212 to run simulations on part twins to predict deformation or failure in an additively manufactured item.

In embodiments, models trained by machine learning system 10210 may be utilized by artificial intelligence system 10212 to run simulations on part twins to optimize the construction process to minimize the occurrence of deformation.

In an embodiment, a model trained by machine learning system 10210 may be used by artificial intelligence system 10212 to run a simulation on a product twin to predict the price of the product. In embodiments, predictions of prices may include: (a) predictions based on market prices for similar items (and/or projections of such prices); (b) forecasts based on forecasted demand; (c) forecasts based on committed demand; (d) predictions based on smart contract terms and conditions; and/or (e) forecasts based on costs, including, inter alia, materials, energy costs, shipping, and labor (which may include a range of profit/marking amounts from base costs to arrive at prices). In embodiments, price predictions may include wholesale prices, retail prices, volume prices, location-based prices, etc.

In embodiments, models trained by machine learning system 10210 may be utilized by artificial intelligence system 10212 to run simulations on part twins, product twins, and printer twins to generate additive manufacturing estimates.

In embodiments, the model trained by machine learning system 10210 is sent to artificial intelligence system 10212 to run simulations on part twins, product twins, and printer twins to generate printing-related recommendations for users of the platform. can be used by In embodiments, recommendations may relate to selection of materials for printing. In embodiments, the recommendation may relate to the selection of an additive manufacturing technology. In embodiments, recommendations may be related to timing of manufacturing.

In embodiments, models trained by machine learning system 10210 are utilized by artificial intelligence system 10212 to run simulations on part twins, product twins, and printer twins to predict delivery times for additive manufacturing operations. It can be. The simulation may include changes in priority levels to determine expected delivery times under different priority levels (such as showing the trade-off between latency and price/cost).

In embodiments, the model trained by machine learning system 10210 is used by artificial intelligence system 10212 to run simulations on part twins, product twins, printer twins, manufacturing node twins, etc. to predict cost overruns in the manufacturing process. It can be used.

In embodiments, models trained by machine learning system 10210 may be used to create part twins, product twins, printer twins, etc. to optimize production sequencing of parts based on quoted price, delivery, sales margin, order size, or similar characteristics. and artificial intelligence system 10212 to run simulations for the manufacturing node twin. In embodiments, optimization may include optimization based on public data, such as market data, website data, manufacturer-provided data (such as by API), and/or terms and conditions of a set of smart contracts related to such characteristics. .

In embodiments, models trained by machine learning system 10210 may be utilized by artificial intelligence system 10212 to run simulations on part twins, product twins, and printer twins to optimize cycle times for manufacturing. there is. In embodiments, optimization of cycle time includes time for post-processing (which can vary dramatically depending on part specifications and additive manufacturing technology).

In embodiments, an instruction set for additive manufacturing uses a blend of natural language-based artificial intelligence and other artificial intelligence to handle and/or generate images and/or spatial representations, e.g., DALL- from OpenAI™. E language models or models for converting images to 3D models and/or other Transformer language models (a combination of text-based and image-based models) further combined with models for converting images or 3D models to an additive manufacturing instruction set. can be automatically generated from text descriptions. A hybrid, transformative artificial intelligence system, for example, generates a set of parameters that represent a set of semantic objects (such as a pair of glasses and a cat) and have cat-like properties (such as whiskers or cat-eye lenses). It can be trained to create output designs (such as eyeglasses) and translate the output designs into a set of additive manufacturing instructions. In such embodiments, a user may, for example, enter a text string for a desired output and be presented with various 3D models representing options. The user can select the desired option and initiate an additive manufacturing operation to create the item. In embodiments, a platform may track interests, attributes, search results, profiles, news topics, or other factors to generate a set of input text strings to generate a set of recommended objects for additive manufacturing for the user. In embodiments, recommendations are based on similarity to other users, such as based on clustering techniques. In an embodiment, recommendations are based on collaborative filtering.

In embodiments, digital twin system 10214 is configured to communicate with a user through multiple communication channels, such as conversation, text, gestures, etc. For example, a digital twin can receive a query from a user regarding a distributed manufacturing network entity, generate a response to the query, and communicate such response to the user. Additionally, digital twins can communicate with each other to learn and identify similar operating patterns and problems in other distributed manufacturing network entities, as well as steps taken to solve those problems. For example, a digital twin of two manufacturing nodes, or a digital twin of a part, a printer, and a manufacturing node can communicate with each other to resolve or respond to customer requests.

118 is a schematic diagram illustrating an example implementation of an autonomous additive manufacturing platform including various components along with other entities in a distributed manufacturing network according to some embodiments of the present disclosure.

Autonomous additive manufacturing platform 10110 may collect data from one or more entities, including users, programs, and data sources 10114. Data acquisition system 10602 within user interface 10112 includes a set of interfaces, such as a chat interface 10604, a smart voice interface 10606, and a file upload interface 10608, to collect data from one or more users of the platform. can do. Additionally, one or more sensors 10610 including a camera and a machine vision system, an acoustic/sound sensor (e.g., optionally having a microphone including multiple microphones in an array), a power and energy sensor, a mass sensor, a location Sensors such as temperature sensors, humidity sensors, pressure sensors, viscosity sensors, flow sensors, chemical/gas sensors, strain gauges, thermal imaging, hyperspectral imaging, sound sensors, air quality sensors, etc. can provide data to the platform (10110). there is. Data source 10114 may also include a program, a feedback source 10612, and a data library 10614 that provide result data from machine learning system 10210.

In embodiments, data visualization 10615 in user interface 10112 may be a dashboard, interface for users of platform 10110 to visualize information related to distributed manufacturing network 10130 or one or more entities within network 10130. and can provide a set of integrations. For example, a dashboard can provide visualizations containing information related to the digital thread for distributed manufacturing network entities, such as 3D printed parts or products. Other dashboards may provide visualizations containing information regarding real-time visibility of the status of manufacturing orders. An alternative dashboard may provide visualizations that include information related to batch traceability to identify parts from the same batch. The dashboard can provide visualization of demand factors including forecasted demand, inventory levels, etc. A search interface may be provided to resolve queries from one or more users based on part, machine, production date, or location. In embodiments, a virtual reality (VR) system may be integrated with a data visualization 10615 and modeling system 10620, thereby allowing users to build 3D models in VR. In embodiments, a virtual reality system allows a user to use scanned data (such as a point cloud) and/or a combination of model-based VR and scans (and/or other augmentations or overlays, such as in augmented reality and/or mixed reality models). It can be integrated with a scanning system 10617, such as allowing to build a model consisting of. This may also involve a broader set of user interactions to develop part designs without in-depth expertise, including using augmented reality (AR) and mixed reality (MR).

In embodiments, user interface 10112 may include a single-click preprocessing process that triggers preset configurations for part orientation, support determination, tool path creation, and/or nesting.

In embodiments, user interface 10112 may include a single-click post-processing process that triggers preset configurations for powder removal, support removal, and surface finishing.

Users of the platform can also use the design and simulation system 10116 to build CAD and STL files that capture the design of the part or product to be printed. A set of design tools 10616 and design libraries 10618 may allow users to build models in modeling system 10620 and run simulations in simulation environment 10622. In embodiments, designs of parts or products may be captured in a variety of file formats, including but not limited to IGES files, SolidWorks files, Catia files, ProE files, 3D Studio files, STEP files, and Rhino files. In embodiments, designs may be in the form of digital images such as PNG files, JPEG files, GIF files, and/or PDF files, as well as scanned data formats such as point clouds generated by laser scanning, and ultrasound, MRI, -Can be captured as output from line, electron beam, radar, IR and other scanning systems.

Data storage system 10204 may store event data 10628 for entities or assets of distributed manufacturing network 10130 over time, including, for example, parts or products or any other assets or entities described in this application. and to maintain records of state data 10630, data may be stored in a distributed ledger 10624, a digital thread 10626, etc.

In an embodiment, digital thread 10626 organizes information related to the entire lifecycle of an item produced by additive manufacturing, such as a part, from design, through modeling, production, verification, use and maintenance, and then disposal.

In embodiments, digital thread 10626 may be associated with one or more additive manufacturing machines, or tools, including post-processing tools such as CNC equipment, robotics support, part/part marking, metrology equipment, etc., across multiple manufacturing facilities/locations. Organize information.

In embodiments, digital thread 10626 optionally includes aggregated, linked, or integrated information from multiple configurations into an overall product digital thread, from design to modeling, production, verification, use, and maintenance. It consists of information related to the entire life cycle of the product, from repair to disposal.

Data processing system 10206 processes data collected by data collection and management system 10202 as detailed in Figures 115, 116, and 118 or elsewhere in this application or in documents incorporated by reference herein. Together, the artificial intelligence system 10212 (including the machine learning system 10210), the digital twin system 10214, and the control system 10216 optimize and adjust process parameters in real time.

Manufacturing workflow management application 10208 may manage various workflows, events, and applications related to production or printing and value chain management. In embodiments, matching system 10632 may help match a set of customer orders with a set of additive manufacturing units 10102 or manufacturing nodes. Orders may include firm orders, conditional orders (e.g., based on price contingencies, timing contingencies, or other factors), aggregated orders, custom orders, volume orders, time-based orders, and the like. In embodiments, orders may be represented as smart contracts, such as operating on a set of blockchains. Matching additive manufacturing capabilities, locations of customers and manufacturing nodes, available capacity at each node, material availability, price (including material, energy, labor and opportunity costs of other available uses for the capacity), and timeline. It can be based on factors such as requirements. In embodiments, different parts of a product may be matched to different manufacturing nodes, and the product may be assembled at one of the nodes, or by a robot (e.g., while in transit, e.g., positioned in a vehicle or shipping container) before finally being delivered to the customer. system) can be assembled elsewhere in the value chain network.

In embodiments, an additive manufacturing platform may be configured to maintain an inventory of available parts for large aircraft or marine systems where redundancy is mandated by customization and/or regulations. In embodiments, example systems include two, three, or more redundancies compared to a basic operating system. In this example, a particular system may utilize off-the-shelf products to fill third, fourth, etc. redundancies when previously a full inventory was required to properly supply the entire third, fourth, etc. redundancy. In light of this disclosure, it will be understood that the present disclosure will not be applicable to all systems in that some critical systems may only accept these components as an additional layer of redundancy over already mandated supplies. During flight, the desire to minimize weight and energy consumption may limit the desire to produce certain parts, but the ability to produce parts for longer endurance flights to accommodate cabin requirements provides some in-flight functionality. It can be one motivation to do it. For example, locking components that may fail during flight, such as latches, hinges, seat belts, etc., can be replaced or temporarily locked closed to improve cabin safety. Components that may become loose can also be shimmed or temporarily held in place by custom printed parts to wedging or holding the part in place through flight. Examples include maintaining avionics components on the dashboard, overhead, or other cockpit controls, maintaining entertainment items in the galley, maintaining seats on seat rails, etc.

In an example, an additive manufacturing platform could be used to create additional inventory to equip an aircraft with configurable items during flight, which would result in a minimum equipment list being required for the flight, before the aircraft lands and returns to the gate for service. Before these parts can be replaced, and thus at least contribute to repairs, which would otherwise not require an early landing, may then prevent the next deployment of the aircraft until its desired use.

In seafaring embodiments, additive manufacturing platforms can be used to create additional inventory to equip sailing vessels with configurable items during a voyage, thereby requiring a mandatory minimum equipment list for embarkation (etc.), and allowing the vessel to Before anchoring and reloading, these parts are replaced and, therefore, at least contribute to repairs, which would not otherwise require diversion and early landing, but may then prevent the next timely deployment of the ship until its next desired use. .

In embodiments, the additive manufacturing platform may be configured to coordinate with land-based additive manufacturing assets to coordinate the configuration of coordinated parts and parts of larger assemblies, so that downtime at ports or hangers can be minimized. In this example, the entity providing timely maintenance inventory may augment one or more offers or coordinate one or more offers within a port or hanger system that may coordinate with one or more field systems active during the voyage and/or flight. By doing so, you can extend your reach and depth.

In embodiments, matching system 10632 helps match additive manufacturing tasks to engineers where matching can be based on factors such as task complexity, engineer experience, and expertise. In embodiments, matching system 10632 helps match additive manufacturing tasks to the location and/or availability of finish operators, where matching may be based on factors such as task complexity, operator experience, and expertise. In an embodiment, matching system 10632 helps match additive manufacturing tasks to sets of additive manufacturing units 10102.

In embodiments, scoring system 10634 helps score and rank various entities within distributed manufacturing network 10130, such as based on performance, quality, timeliness, condition, status, etc. In embodiments, scoring system 10634 helps rank manufacturing nodes based on customer satisfaction scores, such as to meet customer requirements. In embodiments, the scoring system 10634 ranks engineers or other workers based on conditions/performance in completing additive manufacturing tasks, including time required, quality of output, energy used, and other factors. helps to do In embodiments, scoring system 10634 may be used to complete additive manufacturing tasks, including process metrics, output metrics, product quality measures, economic measures (such as ROI, yield, profit, etc.), customer satisfaction measures, environmental quality measures, etc. Assists in grading additive manufacturing units 10102 based on their condition or performance.

In an embodiment, order tracking system 10636 helps track product orders through their movement in distributed manufacturing network 10130 until they are ultimately delivered to the customer. Order tracking system 10636 may receive status data from various entities in distributed manufacturing network 10130 on a real-time or near real-time basis. For example, a 3D printer can provide updates on production step data or a delivery system can provide updates on product location. This information can then be tracked via order tracking system 10636 on a real-time or near-real-time basis, such as by user or customer identity. Workflow manager 10638 manages the complete 3D printing production workflow for distributed manufacturing network 10130, including various events, activities, and transactions related to one or more entities in network 10130.

In an embodiment, alert and notification system 10640 provides alerts, notifications, or reports regarding one or more events to users or customers of network 10130. For example, alert and notification system 10640 may receive data related to specific production parameters or errors based on monitoring of the production workflow for which alerts and notifications may be generated. Such alerts, notifications, or reports may subsequently be sent to the user or customer's computing device (e.g., via email, text message, instant message, phone call, and/or other communications (e.g., using the Internet or other data or messaging networks)). For example, it can be transmitted to a computer, tablet computer, smartphone, phone, mobile phone, PDA, TV, game console, etc.).

In embodiments, error notifications may provide options for use of platform 10110 related to continuing or stopping production or making adjustments to design or production settings.

In another example, users or customers of the distributed manufacturing network may be provided with customized reporting, including live status and analytics, based on real-time and historical data from the distributed manufacturing network 10130. In embodiments, custom reporting may include data and analytics related to demand, production capacity, material usage, workflow inefficiencies, output types, output parameters, materials used, costs, ROI, etc. across one or more manufacturing nodes within the network. there is.

In an embodiment, payment gateway 10642 manages the entire billing, payment, and invoicing process for customers ordering products using distributed manufacturing network 10130. This may include recording an event or transaction on an account or ledger, such as a distributed ledger, such as a blockchain-based ledger. Payments may be allocated according to a set of rules for allocating payments across payers, eg embodied in a smart contract; For example, copyright protection or printing from another proprietary instruction set may trigger royalty payments to intellectual property owners, administrators, etc.

It will be apparent that these applications provided by platform 10110 are presented by way of example only and should not be construed as limiting in scope and that many other applications may be provided for managing one or more aspects of distributed manufacturing network 10130. something to do.

In embodiments, an authentication application may be provided to authenticate the identity of a user of the platform through one or more authentication mechanisms, including a simple username/password mechanism, a biometric mechanism, or a cryptographic key exchange mechanism. Similarly, an authorization application may define the roles and access rights of users of the platform such that users with different roles are provided with different access rights. For example, “administrator” or “host” privileges may allow a user of the platform to change platform configuration, add and remove programs, access arbitrary files, and manage other users on the platform; “Engineer” privileges may allow users of the platform to operate the platform; And “service” privileges may allow users of the platform to access a subset of administrator privileges to perform maintenance and repair activities.

Some other example applications provided by platform 10110 for production management include part marking, slicing tool selection, alerts and notifications for feedstock supply, printing queue management, printer floor management, and job scheduling (across multiple units). including), finishing work management, packaging management, logistics preparation, etc. Some example applications provided by platform 10110 for production reporting include order failure reporting, management information system alerting, remote quality assurance, authentication, indexing, and the like. Some example applications provided by platform 10110 for production analytics include order matching, production failure analysis, warranty management, etc. Some example applications provided by platform 10110 for value chain management include payment processors, digital format conversion, production restrictions, export restriction filtering, etc.

In an embodiment, the platform 10110 includes an Enterprise Resource Planning (ERP) system 10644, a Manufacturing Execution system (MES) 10646, a Product Lifecycle Management (PLM) system 10648, and a maintenance management system (MMS) 10650. , Quality Management system (QMS) (10652), certification system (10654), compliance system (10656), robot/cobot system (10658), SCCG system (10660), etc., which are incorporated by reference into this application. Integrates with one or more third party systems of various types described in the documentation. In embodiments, the platform is integrated into a value chain network control tower system, such as to manage a set of value chain network entities.

In embodiments, the API system facilitates the transfer of data between platform 10110 and one or more third-party systems. API systems are used for conveying command sets, for conveying alerts, notifications, etc., for sending event streams (such as workflow-related events), for sensor data (such as process sensing from manufacturing, environmental sensing, etc.) For communicating data, handling user data, processing payments, integrating with smart contracts, blockchain and other systems, communicating data with AI systems, 3D rendering and It may consist of a set of APIs, including those for transferring data to and from other modeling systems.

In an embodiment, the enterprise resource planning (ERP) system 10644 performs functions such as finance, sales, marketing, service, engineering, product, within a manufacturing node and across multiple manufacturing nodes within the distributed manufacturing network 10130, among other functions. We help streamline and integrate business processes across administration, accounting, procurement, distribution, resources, project management, risk management and compliance. An ERP system 10644 may tie together various production and value chain processes in a distributed manufacturing network 10130 and enable the flow of data between them.

In an embodiment, a manufacturing execution system (MES) 10646 connects machines, processes, equipment, tooling, and materials to streamline manufacturing operations within a manufacturing node and across multiple manufacturing nodes in a distributed manufacturing network 10130. Monitor. MES (10646) can integrate processes across production, distribution, supply chain, maintenance, quality and labor operations. Additionally, MES 10646 may coordinate with other systems and entities within distributed manufacturing network 10130 to help make decisions related to advance planning, production capacity analysis, inventory turns, and lead times.

In embodiments, an additive manufacturing platform, such as that associated with a value chain or other network, may be designed, prepared, configured, and/or deployed to support the design, development, manufacturing, and distribution of: trucks, trains, airplanes, boats, Parts and maintenance materials (e.g., oil, gas, other chemicals) for vehicles used to distribute products, which may include drones, etc.; Parts and maintenance materials for machines (e.g. robots) used in packaging products; Parts and maintenance materials for tools and machines (e.g., robots) used to move packaged products from warehouses to vehicles; Parts repairs for existing (and in-service) parts; Missing parts or parts from otherwise off-the-shelf products and some other parts or components for the design, development, manufacture and distribution of maintenance materials.

In embodiments, an additive manufacturing platform as described herein may be designed, prepared, configured, and/or positioned to support monitoring of packaging materials (e.g., boxes, boxes, wrap materials, etc.) “There is a need to create more.” Additive manufacturing platforms can address “recall” situations by adding or modifying products in the warehouse, monitoring problems with vehicles, machines, tools, and other equipment used, and then “receiving” needed parts or materials. Tools can be replaced “accordingly,” and tools can be created on-demand as needed by workers or robots within the warehouse/distribution network.

In embodiments, an additive manufacturing platform as described herein processes manufacturing inputs, e.g., using an artificial intelligence system (e.g., a robotic process automation system trained on a training set of professional service visit data). Designed, prepared, configured and/or arranged to assist in determining a recommended action, which in embodiments may involve replacement of a part and/or repair of a part, or some other activity. In embodiments, an additive manufacturing platform may automatically determine that an element should be additively manufactured to facilitate repair, such as when a complementary component can be created to replace a worn or missing element. In an exemplary embodiment, some techniques and/or techniques that may be utilized in conjunction with a warehouse/distribution center may include, but are not limited to: transporting different tools, parts, and/or packaging in real time (i.e. , providing and/or including a number of source materials to create (on the spot); Use AI to design products, configure manufacturing processes (including the packaging material creation process), schedule, prioritize, and/or logistics tasks (including parts, materials, etc.) without disrupting other typical processes involved in a warehouse/distribution center. optimizing the efficiency of warehouse processes to replace; Enriching the AI with input/source/training set data related to design factors, economic factors, quality factors, etc. involved in certain example embodiments (in particular, e.g., manufacturing processes of component materials required for machinery and/or packaging materials) using sensors and monitoring of data to adjust); Input, process data to run simulations of individual processes or combinations of processes to anticipate material needs to be able to produce or manufacture tools, parts, packaging, and/or fixture machinery in real time (as required) and combining the output with a digital twin; Additive manufacturing nodes within a mesh and/or fleet for coordinated operation within a warehouse/distribution network in an efficient manner for producing tools, parts, packaging, and/or other materials used to fasten machines in real time. networking; Robots that can be attached to a machine and then print directly on the product, printing tools, printed parts for machines used in warehouse/distribution networks, printed packaging, and/or printing materials used to hold machines in place in real time. to use; Enumerated within a warehouse/distribution center network process to secure products, produce tools, produce parts, produce packaging, and/or produce other materials to secure machines (as needed) in real time. Using a hybrid/pair of different types of 3D printing additive manufacturing, including any and all items created.

In embodiments, a product lifecycle management (PLM) system 10648 helps manage parts or products throughout their entire lifecycle, from concept and design through manufacturing and distribution to customer use and service. The PLM system 10648 can contain accurate, real-time product information across the lifecycle and value chain. This involves developing a product in response to feedback from one or more distributed manufacturing network entities such as customers, distributors, logistics providers, regulators, safety experts, service specialists, salespeople, product managers, designers, resellers, etc. who use the product. and help manage it. This can also enable rapid customization of products and accelerated proof of concept during the product development stage. Additionally, it can help predict product demand and pricing, improve customer engagement, conduct product testing during customer use, and provide proactive warranty management.

In an embodiment, Maintenance Management System (MMS) 10650 monitors a set of 3D printers, cutting tools, filters, machine lasers, and other machines, manages spare parts, maintains records, and supports artificial intelligence and machine tools. It uses learning models to efficiently self-diagnose maintenance requirements and generate work orders. In embodiments, MMS 10650 monitors, maintains records, and creates sets of other machines, equipment, products, fixtures, or other assets (e.g., spare parts, tools, workpieces, accessories, replacement elements, etc. (to) manage maintenance operations for that set of items, including coordinating the additive manufacturing workflow with other maintenance workflows. In embodiments, this is done in conjunction with automation, such as robotic process automation, such as when an RPA agent is trained on a set of expert interactions to perform or support operations performed by maintenance workers.

In an embodiment, a quality management system (QMS) 10652 determines whether a printed part was produced correctly by comparing real-time sensor data to expected feedback data, where the expected feedback data includes historical data, test data, and machine learning data. Created from at least one In an embodiment, QMS 10652 also generates a warranty certification that includes the duration and scope of coverage of the parts warranty upon determining completion of testing and quality assurance.

In an embodiment, QMS 10652 includes automated part metrology and utilizes a vision system with variable focus optical system and artificial intelligence-based pattern recognition for automated part metrology. In embodiments, a vision system dynamically learns about a conformal variable focus liquid lens assembly and training data of results, parameters and data collected from the conformable variable focus liquid lens assembly to train an artificial intelligence system to recognize objects. It may include a processing system that does. The adaptive variable focus liquid lens assembly can continuously adjust based on environmental factors and based on feedback from the processing system to generate training data that corresponds to the physical light that the image represents while providing greater context. By training the vision system to recognize objects using variable optical parameters through a liquid lens assembly, the processing system can learn about the most optimal optical settings for detecting the object. Much more dynamic input to the vision system can result in creating richer context and providing superior object recognition.

In an embodiment, certification system 10654 is configured to generate workflow and process control documentation to obtain a certificate of conformity from one or more manufacturing certification bodies or standards bodies. In embodiments, one or more manufacturing certification bodies or standards bodies include the International Organization for Standardization (ISO), a European Certification (CE marking) body, Underwriters Laboratories (UL), the Society of Automotive Engineers (SAE), the Federal Aviation Administration (FAA), Includes TUV SUD, DNV GL, AS9100, IAQG 9100, American Society of Testing and Materials (ASTM), NIST (Research, Measurement Sciences, and Standards), Fraunhofer Institute (Research), and Sandia National Labs (Research).

In an embodiment, compliance system 10656 is configured to perform compliance checks on 3D printed parts. In embodiments, compliance checking may be performed by or from robotic process automation, such as when a compliance model or algorithm is trained by a qualified expert in certification/compliance to specific requirements on a training set of compliance review data, etc. It is achieved with the support of In embodiments, a domain-specific or subject-specific model, such as for each compliance domain or subject, such as compliance with environmental standards, material standards, structural standards, chemical standards, safety standards, electrical standards, fire-related standards, etc. A set of can be trained.

In embodiments, robot/cobot system 10658 may include an autonomous robotic system or arm unit integrated with a set of additive manufacturing units 10102. For example, the additive manufacturing unit 10102 may be included within the housing or body of a robotic system, such as a multipurpose/general purpose robotic system such as one that simulates human or other animal species capabilities. Alternatively or additionally, the additive manufacturing unit 10102 may be configured to deliver additive deposition from a nozzle disposed on a working end of a robot arm or other assembly.

In embodiments, autonomous additive manufacturing platform 10110 may create and manage profiles of different distributed manufacturing network entities. For example, a profile may include, without limitation: A part or component profile with an accompanying part data structure may store part-related information and component-related information, such as name, number, class, type, material ( s), size, shape, function, performance specifications, etc.; a batch profile with an accompanying batch data structure for storing batch-related information including the batch number; Bundle date, bin number, bundle type, location information (e.g., point of origin), bundle inspection data, etc.; Machine Profile, which involves machine data structures for storing machine-related information, including identifiers, names, classes, functions, etc.; a manufacturing node profile with an accompanying manufacturing node data structure for storing information related to the manufacturing node including identifier, location, order history, production capacity, and previous product designs; a packager profile with accompanying data structures for storing packaging-related information; a user profile, which involves a user data structure for storing user-related information; In particular, it includes a behavior profile with accompanying data structures for storing behavior information. Some examples of users of platform 10110 include designers seeking to create designs for manufacturing; Engineers looking to print and manufacture parts; CFO trying to optimize prices for production; Alternatively, it may include customers who wish to print products. Users may use role-based users, such as executives and other role-based digital twins, as described in connection with other use cases referenced in this application and the documents incorporated herein by reference. It can include many different users, such as various users, consumers of automatically generated data stories, etc.

The metal additive manufacturing platform 10110 described in this application can help automate and optimize a very wide range of manufacturing and value chain functions. Some examples of such functions include process and material selection, feedback creation, design optimization, risk prediction and management, sales and marketing, supply chain for manufactured products and/or related items or services (e.g., parts, accessories, etc.), among others. and coordination with logistics workflows (including reverse logistics and returns), maintenance workflows, recycling workflows and customer service. 119 automates manufacturing functions and subprocesses, including process and material selection, hybrid part workflow, feedstock composition, part design optimization, risk prediction and management, marketing and customer service, according to some embodiments of the present disclosure. A schematic diagram illustrating an example implementation of a platform for management 10110.

Process and material selection

The selection and use of one or more processes or materials for additive manufacturing can be automated and optimized. Platform 10110 can take product requirements in terms of part characteristics, price, performance characteristics, etc. as input and automatically determine a process or material to build the part. The artificial intelligence system 10212 is capable of analyzing materials, including their structure, stresses, deformations, wear, load bearing, response to contamination, chemical interactions with other materials, interactions with biological factors (antibacterial, antiviral, toxic), etc. Model information may be consumed, including physical, chemical and/or biological models of behavior. Artificial intelligence system 10212 may then automate and optimize process and material selection, including based on expert feedback and/or feedback from testing/results.

Referring now to Figures 115, 116 and 119, example embodiments for automating process and material selection are described.

The part design, including model information and product requirements, is presented to design and simulation 10116, where it is evaluated for manufacturing compatibility with at least one type of additive manufacturing unit 10102 within manufacturing node 10100. Design and simulation 10116 includes artificial intelligence 10212, simulation management 10514, printer twin 10506 (which in embodiments may be a twin of any type of additive manufacturing unit), and processes and materials to perform optimization. May be assisted by selection twin 10702. An example analysis includes simulating part design dimensions and accuracy using printer twin 10506 in digital twin system 10214 and comparing to available 3D printer work envelopes and specifications.

After the part design is verified to be compatible with one or more of the additive manufacturing units 10102 within manufacturing node 10100, the part data for manufacturing may be optimized for export in design and simulation 10116. For example, an optimized STL file can be created from a finely meshed 3D CAD surface model to meet part accuracy requirements and then exported to an autonomous additive manufacturing platform 10110.

Autonomous additive manufacturing platform 10110 may include a process and material selection system 10704. Using optimized part data from design and simulation 10116, external information, including price and market-related information from sources such as value chain entities 10126, and assistance from artificial intelligence systems 10212, the process and material selection system 10702 performs analysis to select one or more of additive manufacturing units 10102 for manufacturing the part. In one example, process and material selection system 10702 can analyze the availability and cost of printer feedstock materials to select an additive manufacturing unit 10102 that manufactures parts according to specifications while optimizing for lowest manufacturing cost.

116, 118, and 120, when manufacturing is completed, part and process data related to the results of the 3D printing process are collected by data collection and management system 10202. The resulting data is provided to the machine learning system 10210 along with simulation, external, and training data to train or improve the initial machine learning model 10213.

The following is an example of autonomous design verification and selection of 3D printing processes and materials. 114 and 115, part design data is entered into user interface 10112 and then provided as input to design and simulation 10116 for part verification. Part design data provided in user interface 10112 may include the following part specifications and ordering requirements: a shape or geometry described by a 3D CAD solid model; Loading use cases that apply to a given 3D CAD model; Component design stress factor of safety: > 2; maximum section weight; Corrosion requirements: Compatibility with sea water and salt spray; Order Part Quantity 10; and delivery time.

With assistance from artificial intelligence system 10212, design and simulation 10116 performs a number of screening analyzes including: a material analysis that identifies titanium, Inconel, and 316 stainless steel as materials that meet corrosion requirements; materials analysis assisted by simulations from printer twin 10506 and process and material selection twin 10702, identifying powder bed fusion or metallic material extrusion as a 3D printing process matching the availability of additive manufacturing units 10102; Stress and weight matrix analysis calculated for part geometry and loading, eliminating Inconel and 316 stainless steel due to weight considerations, but qualifying titanium for both weight and maximum stress. Following completion of the screening analysis, process and selection system 10704 is used to complete selection of the final additive manufacturing unit 10102 from the subset of additive manufacturing units 10102 that can be used for manufacturing.

HYBRID PART WORKFLOWS

The selection and use of one or more hybrid manufacturing workflows optimized for applying additive materials onto existing parts can be automated to create modified part assemblies. Hybrid part workflows can be used to develop new manufacturing processes, repair existing parts, and modify existing parts to improve value chain outcomes.

Autonomous additive manufacturing platform 10110 may take as input existing and OEM part information, including physical, chemical, manufacturer specifications, etc., including information based on expert feedback and/or feedback from testing/results. AI system 10212 uses the input data to assist in automatic validation of portions of one or more hybrid workflows in workflow management application 10208.

In the component repair example, data from user interface 10112 and data source 10114 is provided to design and simulation 10116. Exemplary data includes a combination of measurements and expert observations and/or OEM part information such as specifications and CAD models. Design and simulation system 10116 analyzes part dimensions and material repair requirements with reference to their compatibility with at least one type of additive manufacturing unit 10102 within manufacturing node 10100. Design and simulation 10116 may be assisted by artificial intelligence 10212, simulation management 10514, and digital twin systems 10214, for example, analysis of modified parts using available 3D printer capabilities. It may include the use of printer twin 10506 and part twin 10504 in digital twin system 10214 to simulate manufacturing results or determine compatibility of available 3D printer materials with OEM part materials.

After the modified part is verified by design and simulation 10116 as compatible with one or more of the additive manufacturing units 10102 within manufacturing node 10100, the modified part data is exported to autonomous additive manufacturing platform 10110. , where process and material selection system 10704 selects one or more of the additive manufacturing units 10102 for manufacturing using one or more hybrid workflows. Exemplary hybrid workflows include rebuilding worn part areas or replacing chipped or cracked areas of the part.

118 and 119, when modified part manufacturing is completed, part and process data related to the results of the 3D printing process are collected by data collection and management system 10202, where modified part parameters, measurements, The data may be exported to systems responsible for managing assurance, safety, and related compliance, such as ERP systems 10644, certification systems 10654, compliance systems 10656, etc. In embodiments, the data may be used to set parameters for a smart contract, such as filling out warranty-related, safety-related, liability-related, or other provisions of the smart contract. The platform and/or smart contract may store data on the blockchain.

In embodiments, a hybrid manufacturing workflow may be used to modify an existing part design to create a new design, for example, when incorporating new functional or safety features that improve part performance.

In embodiments, hybrid manufacturing workflows may be used to create new parts comprising multiple materials, which may require more than one 3D printer or 3D printing process to create targeted parts or product characteristics.

114 and 115, in an embodiment, a hybrid manufacturing workflow may specify and manage specialized pre-processing (10104) and post-processing (10106) for manufacturing an additive manufacturing unit (10102). Examples include part cleaning, machining, grinding, surface finishing, etc. to enable 3D printing or to create modified parts that meet the original equipment part specifications.

FEEDSTOCK FORMULATION

The selection, purchasing, and management of 3D printer feedstock can be automated and optimized to improve manufacturing efficiency, control supply chain logistics and costs, and provide new part production capabilities.

Referring now to FIG. 119 , feedstock composition system 10706, assisted by artificial intelligence 10212 and feedback composition twin 10708, can configure the 3D printer according to production requirements, supply chain conditions, pricing and availability information, or other data. Automatically formulates and adjusts feedstock. For example, the feedstock composition system 10706 can select a commercially available feedstock, such as Ni alloy 718 from GE Additive, or an equivalent at lower cost from a commercially available elemental material. Local manufacturing of materials can be proposed. In embodiments, price and availability information may be provided, e.g., by an API of the platform and/or feedstock composition system, a smart contract that provides current and/or future (e.g., in a spot market at a designed time in the future) price information. A set of terms and conditions for a set of smart contracts, such as a set of terms, material type, material quality (e.g., if there are different grades of material that can be purchased as feedstock), or other characteristics (e.g., material origin). ) (e.g., recycled or recycled from other sustainable sources, mined with sustainable practices, purchased from ethical sources, etc.) It can be managed by processing (including by time and delivery location). In embodiments, the platform may aggregate availability information, pricing, etc. across multiple smart contracts or a mix of smart contracts and other sources (e.g., data inputs and/or offers placed on the platform by API), e.g. Where feedstock may originate in lots or batches from different suppliers, origin locations, etc., the system can provide an aggregated feedstock availability data structure on which to act. The platform may automatically generate a feedstock purchasing plan that may include a set of current purchases, options or future purchases, and plans for future purchases. In embodiments, the platform may be based on changes in conditions such as demand (e.g., when production varies according to plan and/or demand varies according to plan), price (of the end product and/or materials), availability, etc. This allows the feedstock purchase plan to be automatically modified. This includes a training set of feedstock purchasing management data, which may use artificial intelligence, such as any of the machine learning or other artificial intelligence techniques described herein, including supervised, semi-supervised, and/or deep learning. This can be achieved by process automation with trained robots. An artificial intelligence system may, for example, operate on a set of smart contracts through an API or other interface and/or provide a set of modified recommendations for execution by a user or a hybrid of a user and an intelligent agent or other artificial intelligence system. Depending on the plan, the set of contract terms and conditions for the purchase of feedstock can be further adjusted.

In an embodiment, the feedstock composition system 10706 includes a machine learning system 10210, an artificial intelligence system 10212, a machine learning model for feedback composition 10213, a simulation management system 10514, and a feedstock composition twin 10708. ) can be used to formulate one or more custom feedstocks. Machine learning system 10210 may train a model using feedstock data, which may be stored in a feedstock datastore, such as a graph DB that organizes different feedstocks according to performance characteristics. Simulation management system 10514 can run simulations using feedstock composition twins 10708 to vary feedstock properties and record the results of each simulation. In embodiments, printer twin 10506 may also be used to simulate and compare future manufacturing results when changing feedstock composition.

116 and 119, feedstock formulation system 10706 operates in conjunction with artificial intelligence system 10212 and machine learning system 10210. A combination of training, manufacturing results, and external data, such as pricing and availability information and expert and customer feedback, is collected in data collection and management system 10202, where it trains or refines an initial machine learning model 10213 to formulate feedback. It is used to

Referring now to FIGS. 114 , 115 , and 119 , in embodiments, feedstock composition system 10706 may include a physical subsystem that is integrated with one or more of a manufacturing node 10100 and an additive manufacturing unit 10102. These physical subsystems of feedback formulation system 10706 may be managed by autonomous additive manufacturing platform 10110. Manufacturing workflow management application 10208 can include an application to route feedstock materials as needed, and data collection and management system 10202 can provide feedstock inventory levels. The feedstock composition system 10706 may include one or more automated production and transportation systems that deliver feedstock materials and perform feedstock material changes to the additive manufacturing unit 10102.

DESIGN OPTIMIZATION

Optimizing part design for use with an additive manufacturing process typically requires specialized software, equipment, training, technical knowledge, and the ability to provide and interpret process data and manufacturing results. Autonomous or guided product design can be used to improve value chain outcomes by using pre-engineered part libraries or expert systems to provide autonomous part design or expert-assisted design optimized for metal additive manufacturing processes. The resulting workflow and process functionality can be further optimized by including constraints or recommendations based on real-time analysis of selected materials or value chain entities providing data on the availability of 3D printers, part costs and delivery times, etc.

118, part design optimization for a 3D printing process can be automated using design and simulation 10116, where part features and/or class criteria are organized into a design library 10618 and the parts are prepared for manufacturing. It can be used to guide or fully automate the design. Part features and classes have their own minimum design criteria imposed by standards, best practices, engineering experts, etc. Examples of component features include self-lubricating bearings made of sintered metal that must meet the chemical, mechanical and other properties found in the ISO 5755 standard, or electric hand tools whose materials must meet the 1000 V electrical insulation standard found in the IEC 60900 standard. do. Examples of component classifications include components for use in explosive atmospheres where construction materials must be non-incendive, or components for medical tools used in surgery where corrosion properties must comply with the ASTM F1089 standard.

115, 116, 118, and 119, in one example embodiment, a request for a new part with specific functionality is received by user interface 10112 and communicated to design and simulation 10116; Here, the design library 10618 is searched for tested and executable 3D printed part models that match the part features. In embodiments, one or more parts from design library 10618 are recommended to a user, such as through interface 10112, as design recommendations or guidance. In embodiments, a design library may also include a finished assembly and a product assembly in which all components within the assembly meet functional or class criteria.

In embodiments, one or more candidate portions are automatically selected by design optimization system 10710. With the assistance of machine learning system 10210 and artificial intelligence system 10212, design optimization system 10710 optimizes the part design and submits it to autonomous additive manufacturing platform 10110 for manufacturing.

In embodiments, design optimization system 10710 may use machine learning models trained by product design experts. In embodiments, design optimization system 10710 may use machine learning models trained using data and results from previous designs.

In embodiments, design optimization system 10710 may use a generative or evolutionary approach to design. The system can start with a design goal and then explore countless variations by adding constraints before choosing a final design based on an evolutionary model. Evolutionary models are based on the principles of natural selection, where the most optimal design is selected from an initial population of potential designs through a series of evolutionary stages. Generative models include models such as DALL-E™ that blend visual and text-based artificial intelligence systems, as well as additional hybrid models to generate visual, 3D, text, color, texture, strength, flexibility, and many other properties. This may involve using specialized artificial intelligence systems to generate variations of each of a large set of features and to generate combinations such as pairs, triplets, and higher-order n-tuples of features. Includes. In embodiments, a generative model represents a combination of properties shared between semantically distinct objects or subjects, such as a cat and a basket, to generate and/or select a set of designs that implement the shared set of properties. You can create and/or select design instances.

In embodiments, the evolutionary model may be based on genetic algorithm (GA), evolutionary strategy (ES) algorithm, evolutionary programming (EP), genetic programming (GP), and other suitable evolutionary algorithms. In embodiments, the evolutionary model may be based on semantic features, based on design constraints (such as an acceptable color palette for the brand), based on physical or functional requirements (such as surveys, engagement tracking, and/or A/B generated by consumer participation (such as testing), outcome-based (such as sales, profits, etc.), cost-based (materials, manufacturing, logistics, etc.), safety or liability-based, regulatory requirements, or Various feedback and filtering functions are available, such as those based on authentication. In an embodiment, feedback on design evolution is taken from a set of smart contracts, such as a set of smart contracts providing various design variations for purchases, reservations, etc. For example, a design may evolve based on the execution of a favorable smart contract, such as when a particular design is booked through a set of smart contracts at a profitable price and in a favorable volume.

In an embodiment, an evolutionary design system coupled to a set of additive manufacturing units 10102 provides a continuous set of products through a smart reservation contract that allows users to reserve units for manufacturing according to the provided design, and thus The capacity of the additive manufacturing system is continuously involved in evolving the design to provide the most advantageous outcome in smart contracts (e.g., based on measures of profitability) and selling products to users who reserve them through smart contracts. do. Smart contract parameters, including price, delivery period, etc., can be automatically adjusted to take into account manufacturing time, logistics factors, etc. The system is not only an e-commerce system such as for providing products in markets, auction sites, mobile applications, etc., but also other environments where purchases are possible such as on-site systems (kiosks), in-game transaction environments, AR/VR environments, smart displays, etc. It can be configured to integrate with

116 and 119, once manufacturing is complete, part and process data related to the results of the 3D printing process are collected by data collection and management system 10202. The resulting data is provided to the machine learning system 10210 as feedback along with simulation, external, and training data to train or improve the learning model 10213.

RISK PREDICTION AND MANAGEMENT

Referring now to FIG. 119 , risk prediction and management system 10712 interfaces with, links to, or integrates artificial intelligence system 10212 with artificial intelligence system 10212. In an exemplary embodiment, risk prediction and management system 10712 may determine, among other risks or liabilities, risks or liabilities relating to the manufacture, delivery, use and/or disposal of parts, products or other items by distributed manufacturing network 10130. It can be configured to predict and manage.

In embodiments, machine learning system 10210 may perform classification related to risk management, including for parts and products manufactured by distributed manufacturing network 10130 and for the systems, workflows, and other activities they involve; Train one or more of the models 10213 that are used by the artificial intelligence system 10212 to make predictions and/or other decisions.

In an example embodiment, model 10213 may be trained to predict the risk of component failure by detecting the condition of the component. Machine learning system 10210 may include historical data on wear and tear during use, historical data on material deterioration under various ambient or environmental conditions, data on defects or defects found during inspection or reported by the customer or others; The model may be trained using the part data and one or more results associated with the part conditions, including training data sets for results such as similar parts, similar materials, etc., and other data sources. Part data may include any attribute or parameter mentioned throughout this disclosure and the documents incorporated by reference in this application, such as part material, part characteristics, date of manufacture, material supplier, part specifications, etc. In this example, the results used to train the machine learning system 10210 to predict failure of risk, liability, and documents incorporated by reference in this disclosure and this application (e.g., physics, chemistry, biology, materials science etc.) may include expected results from models, such as various types of scientific models, economic models, etc., described throughout, which in embodiments may include, for example, during or after a simulation of a set of events, passage of time, etc. A part twin 10504, product twin, or other twin may be embedded in a digital twin system to model whether it is in favorable operating conditions. In this example, one or more characteristics of part twin 10504 may be changed for different simulations and the results of each simulation may be recorded. Other examples of trained risk prediction and management models may include a model 10213 that is trained to optimize product safety, a model that is trained to identify parts with a high probability of failure, and the like.

In an exemplary embodiment, model 10213 may be used to address, for example, equipment failures, strikes and other labor disruptions, border control activities (e.g., customs inspections, travel bans, etc.), restrictions on shipping, traffic congestion, power outages, storms, and other conditions. Such as natural disasters, disasters, economic disruptions (such as major changes in tariffs), regulatory changes (such as bans on imports or exports or changes in where products can legally be sold or used), pandemics, political unrest, etc. It can be trained to predict the risk that products will not reach customers due to supply chain and other disruptions, such as those caused by various external events. In this example, the model can be accessed from social media feeds, weather patterns, news feeds, websites (e.g. websites providing content related to the above, marketplace websites, research websites, etc.), crowdsourcing systems (economic factors, behavior one or more external sources, such as posing a query or project to the crowd to solicit input on specific factors such as factors, trends, etc.), algorithms (e.g. trained to provide specific predictions of events), etc. Can be trained to predict supply chain disruptions by discovering, extracting, transforming, normalizing, processing, and/or analyzing data from The artificial intelligence system 10212 can then predict and evaluate the impact of the predicted disruption to determine whether supply chain redesign may be required to minimize the disruption. Impact assessment and/or forecasting may use a set of economic, financial or operational models, among others, for example to evaluate primary, secondary and other effects on the overall workflow or system. For example, assessing or predicting the impact of a component's absence on the ability to deliver the system in a timely manner; the impact of reduced or late supplies on sales (for example, missing seasonal windows that have a major impact on product demand for some products, such as Halloween costumes or beach chairs); the impact of reduced or late supply on prices (for example, where anticipated shortages may dictate the need for price increases and/or purchasing restrictions to balance supply and demand and avoid shortages or disruptions or products); Impact on contractual liabilities (e.g. liability for failure to deliver, including the buyer's obligation to pay for costs covered in the market by purchasing substitute items); Impact on brand or reputation; etc.

In embodiments, artificial intelligence system 10212 may utilize environment twin 10714, manufacturing node twin 10510, and/or other twins to run a set of simulations to evaluate the impact of an outage on one or more manufacturing nodes. . Risk prediction and management system 10712 may then initiate a supply chain redesign or product resupply event to minimize the impact of the disruption. Additionally, the results of these events (e.g., improved lead times) may be reported to machine learning system 10210 to strengthen the models used to make decisions.

MARKETING AND CUSTOMER SERVICE

Referring now to FIG. 119 , marketing and customer service system 10716 interfaces with, links to, or integrates with artificial intelligence system 10212. In an example embodiment, marketing and customer service system 10716 may be configured to provide personalized sales, marketing, advertising, promotions, and/or customer service for products or other items offered by distributed manufacturing network 10130. You can.

In embodiments, machine learning system 10210 may be used to classify, predict, and/or make other decisions related to sales, marketing, advertising, promotion, and/or customer service for products manufactured by distributed manufacturing network 10130. train one or more of the models 10213 to be used by the artificial intelligence system 10212.

In an example embodiment, model 10213 may be trained to predict the behavior and purchasing patterns of one or more customers to provide personalized sales, marketing, advertising, promotions, and/or customer service. In embodiments, machine learning system 10210 uses one or more results and customer data associated with customer responses to personalized campaigns, such as search engines, news sites, websites, behavioral analysis systems and algorithms, and consumer sentiment measurement. , models can be trained using a variety of data sources that provide insights into consumer sentiment, behavior, and more, including microeconomic and macroeconomic measures. Models can be seeded with a variety of economic, behavioral, and other models, including demographic, psychological, economic, game theoretic, cognitive, and other models. Customer data includes identity data, transaction and payment data, location data, demographic data, psychographic data, location data, wealth data, income data, sentiment data, loyalty data, loyalty program data, clickstream data (social media, application, (including interactions with websites, mobile devices, AR/VR systems, video games, entertainment content, and other digital content), point-of-sale data, in-store behavior data (e.g., in-store path tracking data, certain types of products) time of stay associated with, etc.), brand loyalty data, shopping data, search engine data (e.g., search topics involving shopping), social media footprint, purchase history, loyalty program data, etc., incorporated herein by reference in this disclosure and this application. It may include any of the types described throughout the included document. Customer twin 10718 can measure outcomes such as customer attention or actions (including mouse movements, mouse clicks, cursor movements, navigation actions, menu selections, etc.) measured through a software interaction observation system, or the purchase of a product by the customer. Tracking may capture a set of customer responses to a marketing or advertising campaign or one or more product recommendations, offers, advertisements or other communications. In this example, one or more parameters of a marketing or advertising campaign can be changed for different simulations of the customer twin, and the results of each simulation can be recorded.

In embodiments, marketing and customer service system 10716 may use artificial intelligence to provide personalized sales, marketing, advertising, promotions, and/or customer service, including delivering personalized marketing and advertising campaigns and providing product recommendations. May interface with system 10212. In embodiments, artificial intelligence system 10212 may utilize one or more of machine learning models 10213 to determine product recommendations. In an embodiment, simulations run by customer twin 10718 may be used to train a product recommendation machine learning model. In each of these examples, the campaign communications, recommendations, etc. may be customized to the customer and represent a product that can be manufactured by an additive manufacturing unit 10102 having a set of attributes that can be delivered to the customer's designated site within a specified time frame at a suggested price. Or it may involve other items. Customization of suggestions/recommendations allows the product to include attributes preferred by the customer, including functional attributes, preferred materials (such as matching those of products already owned by the customer), preferred colors, preferred shapes, etc. Or it may include providing a design of the part. In embodiments, customization may be based, for example, on purchase history information, such as compatible colors, shapes, sizes, material types, connectivity (e.g., to operate as part of a set of connected products), communication protocols, logos, etc. By recommending a product that has, etc., the customer may reference an understanding of products that the customer already owns, for example, if the recommended product can be configured to operate as part of a family of products.

In embodiments, an additive manufacturing platform 10110, such as that associated with a value chain network, is prepared to support the printing of personalized entertainment props, backdrops, and other items in theme parks, cruise ships, theaters and movie productions, and/or other entertainment venues. , may be configured and/or deployed. For example, in the context of a cruise ship, additive manufacturing unit 10102 may be designated to support printing of guest rooms, themed rooms, or furniture based on and tailored to a given theme. Customers can provide preferences in terms of room layout and design, furniture and accessories, which can be printed dynamically. Similarly, for theme parks, additive manufacturing unit 10102 may be designated to support printing of rock walks, rides and other attractions, and for theater and film productions, movie props, costumes, sets, artifacts and other Accessories can be custom printed.

In embodiments, the platform may include input indicating the item being printed (e.g., technical specifications, CAD design, etc.); Input indicating requirements (e.g., need to improve existing roller coaster personnel with custom rockwork, need to build dinosaur replicas, etc.); and inputs captured by cameras, microphones, data collectors, sensors, and other information sources associated with the entertainment venue.

In embodiments that recommend or configure instructions for additive manufacturing, platform 10110 discovers available materials, including fabrics, metals, plastics, etc., configures instructions, initiates additive manufacturing, and determines when the elements are ready for use. Updates can be provided to the owners of entertainment venues, such as updates on what will happen next. Platform 10110 may, in some such embodiments, use artificial intelligence systems 10212, such as trained on expert data sets, to ensure that suitable items are readily available and/or for additive manufacturing to create the item(s). It can automatically determine whether using the system can reduce delays or save costs.

In embodiments, platform 10110 can monitor printing tasks, e.g., through trained AI agents, to determine what other tasks are being done (e.g., to allow proper sequencing of additive manufacturing outputs aligned with the overall workflow). A set of additive manufacturing units 10102 while recognizing the status of other related entities involved in other workflows, such as their priority (e.g., whether it is related to a movie scene being shot), cost of downtime, or other factors. You can automatically organize and schedule sets of tasks across In embodiments, optimization of workflow across a set of additive manufacturing entities may be achieved by having the artificial intelligence system 10212 perform a set of simulations, such as simulations involving alternative scheduling sequences, design configurations, alternative output types, etc. You can. In embodiments, simulations may be performed on additive manufacturing and other manufacturing entities (e.g., those performing cutting, dyeing, etc.), including handoffs between sets of different manufacturing entity types, such as when handoffs are handled by a robotic handling system. may include sequences involving cutting manufacturing entities and/or finishing entities that sew, construct, add customer initials, etc. In an embodiment, a set of digital twins may represent the properties and capabilities of various manufacturing systems, various handling systems (robotic systems, arms, conveyors, etc., as well as human workforce) and/or the surrounding environment.

It will be clear that the above decisions regarding prediction, optimization using artificial intelligence system 10212 of platform 10110 are presented as examples only and should not be construed as limiting. decisions related to forecasting and optimization of prices by CFO Twin (10720); Decisions regarding new product launches by CEO Twins based on behavioral patterns and market trends; There may be multiple use cases including:

In embodiments, autonomous additive manufacturing platform 10110 enables distributed manufacturing network 10130 by managing production workflows within and across one or more manufacturing nodes, thereby enabling manufacturing through sharing of resources, capabilities, and intelligence. Facilitates cooperation across nodes. In embodiments, manufacturing nodes may collaborate to anticipate and predict material supply and product demand. In embodiments, manufacturing nodes may collaborate for design and product development. In embodiments, manufacturing nodes may collaborate to manufacture and assemble one or more parts of a product. In embodiments, manufacturing nodes may collaborate for distribution and delivery of manufactured products.

Accordingly, distributed manufacturing network 10130 may provide “manufacturing as a service” by leveraging the unutilized capacity of one or more 3D printers by exposing the capacity to one or more users/designers seeking to manufacture 3D printed parts.

In an embodiment, a method for facilitating manufacturing and delivery of a 3D printed product to a customer using one or more manufacturing nodes in distributed manufacturing network 10130 includes receiving one or more product requirements from a customer; determining one or more manufacturing nodes, processes, and materials based on product requirements; generating a quote including pricing and delivery timeline; And if the customer accepts the quotation, manufacturing the 3D printed product and delivering it to the customer.

In embodiments, the product requirements may be a 3D printing instruction set that includes a file (e.g., a CAD file and/or an STL file) and any accompanying instructions for printing the product defined in the file.

In embodiments, a distributed manufacturing network may be implemented through a distributed ledger system integrated with a digital thread to store a set of entities, activities, and transactions associated with the distributed manufacturing network.

In embodiments, a smart contract system may communicate with a distributed ledger system and may be configured to implement and manage smart contracts via the distributed ledger. Smart contracts can be stored in a distributed ledger and can contain triggering events. A smart contract can be configured to perform smart contract actions in response to the occurrence of a triggering event. A distributed manufacturing network may be configured to receive instances of a 3D printing instruction set from a user. The 3D printing instruction set may be tokenized so that instances of the 3D printing instruction set can be manipulated as tokens on a distributed ledger. Tokenized 3D printing instruction sets can be stored through a distributed ledger. Commitments from various parties (distributed manufacturing network entities) to smart contracts can be processed. The use of smart contracts in distributed manufacturing networks helps automate distributed manufacturing workflows.

In embodiments, a distributed manufacturing network facilitates the creation of a distributed manufacturing marketplace or exchange for the purchase and sale of additive manufacturing components, products, and instruction sets with manufacturing nodes constituting sellers and customers constituting buyers.

In embodiments, a distributed manufacturing network facilitates the creation of a data marketplace for selling operational additive manufacturing data by manufacturing nodes to a data aggregator. In an embodiment, a data marketplace is built on a distributed ledger and manufacturing nodes are rewarded using digital tokens through smart contracts. In embodiments, the data is anonymized to hide the identity of the manufacturing node that owns the data.

120 is a schematic diagram of a distributed manufacturing network built on a distributed ledger system and enabled by an autonomous additive manufacturing platform, according to some embodiments of the present disclosure.

Distributed manufacturing network 10130 is implemented as a distributed ledger system, where the distributed ledger may be at least partially distributed across nodes of distributed manufacturing network 10139 and may contain blocks that are cryptographically linked. The distributed ledger system stores data related to sets of entities, activities, and transactions in the distributed manufacturing network 10130.

The different manufacturing nodes 10100, 10128, 10800, and 10802 each represent nodes within the distributed manufacturing network 10130. Additionally, an additive manufacturing unit 10102, a pre-processing system 10104, a post-processing system 10106, a material handling system 10108, an autonomous additive manufacturing platform 10110, a user interface 10112, a data source 10114, and Different systems within the manufacturing node, including design and simulation systems 10116, referred to as distributed manufacturing network entities, constitute distributed computing nodes of the distributed ledger system.

A distributed computing node is essentially a computing device having a processor and a computer-readable medium with machine-readable instructions stored thereon, and containing a complete copy of the transaction history of the distributed ledger. Nodes in the distributed ledger can be implemented in a variety of computing systems, including additive manufacturing systems, enterprise systems, inventory management systems, packaging systems, shipping and/or delivery tracking systems, SKU databases, smart factories, etc. Whenever an additional transaction is proposed to be added to the distributed ledger, one or more of the nodes typically verify the proposed additional transaction record, such as through a consensus algorithm. Typically, once a proposed transaction has been verified, for example through some consensus algorithm, the proposed transaction is added to each copy of the distributed ledger across all nodes.

In embodiments, transaction data is verified by nodes through a proof-of-work (POW) consensus algorithm and hashed into an ongoing chain of cryptographically approved blocks of transaction records that constitute a distributed ledger.

In an embodiment, the proof of operation algorithm requires a node to perform a series of computations to solve a cryptographic puzzle. For example, to verify a pending data record, a node may be required to compute a hash via a hash algorithm (e.g., SHA256) that satisfies certain conditions set by the system. Computing a hash in this manner may be referred to in this application as “mining,” and the node performing mining may be referred to as a “miner” or “miner node.” A distributed ledger may, for example, require that the value of a hash be below a certain threshold. In this embodiment, the node inputs a “base string” (i.e., a combination of various types of metadata within the block header, e.g., cause hash, hash of the previous block, timestamp, etc.) into the POW algorithm. You can create a hash by combining it with a "nonce" (e.g., an integer value). In an example embodiment, the nonce may initially be set to 0 when calculating the hash value using the POW algorithm. The nonce can then be incremented by a value of 1, until the node can determine the nonce value that results in a hash value below a specified threshold (e.g., the requirement that the resulting hash begins with a specified number of zeros). It can be used to calculate a new hash value as needed. The first node to identify a valid nonce may broadcast the solution (in this example, the nonce value) to other nodes in the distributed ledger for verification. Once other nodes verify the “winning” node’s solution, the pending transaction record can be attached to the final block in the distributed ledger. In some cases, divergence in distributed ledger copies may occur when multiple nodes compute valid solutions in a short time frame. In such cases, nodes using the POW algorithm accept the longest chain of blocks (i.e. the chain with the largest proof-of-work) as the “true” version of the distributed ledger. Subsequently, all nodes with divergent versions of the distributed ledger can adjust their copy of the ledger to match the true version as determined by the consensus algorithm.

In another embodiment, the consensus algorithm may be a “proof of stake” (“PoS”) algorithm, where verification of pending transaction records relies on the user’s “stake” in the distributed ledger. For example, a user's “stake” may depend on a digital currency within a distributed ledger or the user's stake in a points system (e.g., cryptocurrency, token system, asset sharing system, reputation point system, etc.). The next block within the distributed ledger can then be determined by the record of pending transactions collecting the largest number of votes. A larger stake (e.g. in a given digital currency or token system) results in a larger number of votes that a user can allocate to a particular pending transaction record, which in turn allows that particular user to create a block on the distributed ledger. increases the chances of doing so. In embodiments, the distributed ledger need not be based on a token or cryptocurrency system, but rather may be secured by, for example, conventional or other security techniques. In embodiments such as those involving digital threads, proof-of-stake may be weighted, such that the product manufacturer's votes, customers' votes, etc. are counted more than any third party's.

In another embodiment, the consensus algorithm may be a “Practical Byzantine Fault Tolerance” (“PBFT”) algorithm in which each node verifies pending transaction records using stored internal state within the node. In particular, users or nodes can submit requests to post records of pending transactions to the distributed ledger. Each node in the distributed ledger can then run a PBFT algorithm using the pending transaction record and each node's internal state to reach a conclusion regarding the validity of the pending transaction record. Upon reaching the above conclusion, each node may submit a vote (e.g., “yes” or “no”) to other nodes in the distributed ledger. A consensus is reached between nodes by considering the total number of votes submitted by the nodes. Subsequently, once a critical number of nodes vote “yes”, the pending transaction record is treated as “valid” and is then attached to the distributed ledger across all nodes.

In embodiments, nodes are paid transaction fees for mining activities. In an embodiment, the distributed ledger is a private and permissioned blockchain controlled by a single entity or a consortium of trusted entities, built using pre-built APIs provided for CORD A, Hyperledger, and Quorum.

In embodiments, the distributed ledger is a public permissionless blockchain built on the Ethereum or Bitcoin blockchain. In embodiments, event data related to the movement of goods through a supply chain in a trade finance network may be tracked using an IoT subsystem.

In embodiments, transaction records stored in a distributed ledger may be hashed, encrypted, or otherwise protected from unauthorized access and may be accessible only using a private key to decrypt the stored information/data.

A blockchain may be a single blockchain configured to store all transactions within it, or it may include multiple blockchains, each blockchain being utilized to store transaction records representing a specific type of transaction. For example, a first blockchain may be configured to store shipping data and supply chain transactions, and a second blockchain may be configured to store financial transactions (e.g., via virtual currency).

In an embodiment, a distributed ledger system includes a decentralized application downloadable by entities within a distributed manufacturing network.

In an embodiment, the distributed ledger system includes a user interface configured to provide a set of integrated views of workflows to a set of entities in a distributed manufacturing network.

In embodiments, the distributed ledger system includes a user interface configured to provide tracking and reporting on the status and movement of products from ordering through manufacturing and assembly to final delivery to a customer.

In an embodiment, the distributed ledger system includes a system for digital rights management of entities within a distributed manufacturing network. In an embodiment, the distributed ledger system stores digital fingerprinting information and other information of documents/files, including creation, modification.

In an embodiment, the distributed ledger system includes cryptocurrency tokens to encourage price creation and transfer prices between entities within a distributed manufacturing network.

In an embodiment, the distributed ledger system includes a system for verifying the experience of a manufacturing node.

In an embodiment, a distributed ledger system includes a system for capturing end-to-end traceability of parts.

In an embodiment, a distributed ledger system includes a system for tracking all transactions, modifications, quality checks, and certifications on the distributed ledger.

In an embodiment, the distributed ledger system includes a system for verifying the capabilities of manufacturing nodes.

In an embodiment, the distributed ledger system includes smart contracts to automate and manage workflow in a distributed manufacturing network.

In an embodiment, the distributed ledger system includes smart contracts to execute purchase orders covering work scope, estimates, timelines, and payment due dates.

In an embodiment, the distributed ledger system includes smart contracts for processing payments by customers upon delivery of products.

In an embodiment, the distributed ledger system includes smart contracts for processing insurance claims for defective products.

In an embodiment, the distributed ledger system includes smart contracts for processing quality assurance claims.

In an embodiment, the distributed ledger system includes smart contracts for automated execution and payment of maintenance.

121 illustrates an example implementation of a distributed manufacturing network where digital thread data is tokenized and stored in a distributed ledger to ensure traceability of printed parts from one or more manufacturing nodes within the network, according to some embodiments of the present disclosure. This is a schematic diagram. Users of the distributed manufacturing network 10130 may provide product requirements in the form of a purchase order or 3D printing instruction set 10902. The 3D printing instruction set 10902 includes key specifications and requirements such as product design, materials for printing, quantity to be printed, price a user is willing to pay for the print, and timeline to complete printing. 3D printing instruction set 10902 may also include one or more files (e.g., CAD files and/or STL files) and any accompanying instructions for printing products defined in the files.

Upon receipt, the 3D printing instruction set 10902 is tokenized and stored in distributed ledger 10624 of autonomous additive manufacturing platform 10110. Basic information in the 3D printing instruction set 10902 is stored in the form of a unique record represented by a block number with an address on a distributed ledger, which in turn is represented by a cryptographic token. The cryptographic token captures the value of basic information in the 3D printing instruction set 10902 as ownership or access to a distributed ledger address and tracks the transfer of such ownership between users of the distributed manufacturing network 10130. For example, in Figure 121, 3D printing instruction set 10902 is a random 256-bit integer A091BC3... It is tokenized in the form of and stored in the distributed ledger (10624) represented by address BC22. As a new block is added to distributed ledger 10624 at node 10128, all copies stored at the various nodes, including manufacturing node 10100, manufacturing node 10800, and manufacturing node 10802, are updated with the new block. . Matching system 10632 within autonomous additive manufacturing platform 10110 may assist in matching a purchase order or 3D printing instruction set 10902 to one or more manufacturing nodes or 3D printers. Matching may be based on factors such as printer capabilities, location of customer and manufacturing nodes, available capacity at each node, price, and timeline requirements. In embodiments, a smart contract may be used to trigger conditional logic embodied in the smart contract, such as, for example, tracking satisfaction of a delivery obligation, releasing an insurance obligation (such as insurance covering the product during delivery), etc. , operates on the ledger. In embodiments, smart contracts may be used by, among other things, tax and customs authorities, credit and debit card issuers, distributors and resellers, recipients of commissions, recipients of royalties, recipients of rebates, credits, etc., shippers/carriers, etc. and to the manufacturer, a financial value may be assigned.

In an embodiment, matching system 10632 matches product parts 10904 and 10910 to manufacturing node 10100 for printing, parts 10906 and 10908 to manufacturing node 10128, and parts 10912. and 10914) may be determined to match manufacturing node 10802. Assembly of all components into the final product can be matched to manufacturing node 10800.

Each part also contains a part specification that includes information such as purchase order identifier (orderID), instruction set identifier (fileID), manufacturing node (manufacturerID), 3D printer (printerID), part number (partID), and material and quantity. It can be tokenized to capture information and stored as a record or block in a distributed ledger. The part can then be linked to the block and tracked using a physical tracker using a unique part number, engraving, RFID tag, barcode or smart label unique to the token. In a similar manner, products assembled from all components may also be tokenized and tracked as they move through the distributed manufacturing network 10130 and through various VCN entities 10126 to the customer.

In embodiments, tokenizing a part, product, or 3D printed instruction set may include: This may include wrapping intellectual property, licensing, ownership, financial, time sharing, rental, lending, use sharing and/or other suitable rights into a token.

In embodiments, distributed manufacturing network 10130 may define permissions and/or operations associated with tokens. For example, a token may allow a tokenized 3D printed instruction set to be viewed, edited, copied, purchased, sold, and/or based on permissions established at the time of tokenization by distributed manufacturing network 10130. Permission may be permitted. In embodiments, distributed manufacturing network 10130 may provide for the orchestration of distributed manufacturing marketplaces or exchanges, such as where a 3D printed instruction set may be processed by a host of distributed manufacturing marketplaces or marketplaces, such as, without limitation, and/or Alternatively, they can be exchanged through tokens that are optionally governed by smart contracts that can be configured by manufacturing nodes. For example, an exchange or marketplace could host the exchange of tokenized 3D printing instruction sets, parts, products, expertise, trade secrets, and insights, where transaction terms include (buy/sell model, auction model, donation). predefined and/or configurable (such as configurable smart contracts that enable various transaction models, including models, reverse auction models, fixed price models, variable price models, conditional price models, etc.), wherein metadata is collected and/or or is expressed in relation to a category of distributed manufacturing market or exchange, where relevant content is presented, including market pricing data, substantive content about additive manufacturing, content about providers, etc. Such an exchange could facilitate the monetization of tokenized 3D printing instruction set knowledge expressed in tokens.

In embodiments, the Distributed Manufacturing Marketplace as described herein may be integrated with or within other exchanges, such as domain-specific exchanges, geography-specific exchanges, etc., wherein the Distributed Manufacturing Marketplace may be a subject matter of other exchanges, such as: It can be configured to deal with: changes in other exchanges in the models and algorithms (e.g. pricing models, prediction models, control systems, etc.) used in the distributed manufacturing market; these include supply, demand, price, volume, operating factors, and taking into account the extent to which they influence other factors; A set of items that can be used by other exchanges through a distributed manufacturing unit (e.g., by providing products that can be exchanged on other exchanges, by providing data sets, analytical measurements, etc. that can inform the behavior of other exchanges, etc.) and /or providing a set of data; Providing resource sharing between distributed manufacturing markets and other exchanges (such as enabling shared computation, shared data storage, shared network resources, shared security resources, shared physical locations, etc.); and/or providing integrated coordination of decentralized manufacturing markets and other exchanges. Utilizing shared resources may include, for example, embedding a set of services from another exchange into one or more additive manufacturing units, so as to result in a hybrid of the additive manufacturing unit and other exchange-enabled units. Other exchanges include product exchanges (e.g. e-commerce markets, auction markets, etc.), stock exchanges, commodity exchanges, derivatives exchanges, futures exchanges, advertising exchanges, energy exchanges, renewable energy credit exchanges, knowledge exchanges, cryptocurrency exchanges, and bond exchanges. Exchange, currency exchange, precious metals exchange, oil exchange, exchange for goods, exchange for services, exchange for legal rights (e.g. intellectual property, real estate, rights of publicity, rights of privacy, etc.), or any of a wide variety of others. It could be. This may include integration by APIs, connectors, ports, brokers, and other interfaces, as well as extraction, transformation, and loading (ETL) technologies, smart contracts, wrappers, containers, or other capabilities.

In embodiments, digital twin system 10214 may be configured to present a simulation of a market, exchange, product, seller, buyer, transaction, or combination thereof via a market digital twin. A digital twin or replica can be a two-dimensional or three-dimensional simulation of a market, exchange, product, seller, buyer, transaction, etc. The digital twin can be viewed on a computer monitor, television screen, three-dimensional display, virtual reality display and/or augmented reality display such as a headset, AR goggles or glasses, etc. The digital twin may be configured to be manipulated by one or more users of autonomous additive manufacturing platform 10110. Manipulation by the user may allow the user to view one or more parts of the digital twin in more or less detail. In embodiments, digital twin system 10214 may be configured such that the digital twin can simulate one or more potential future states of a market, exchange, product, seller, buyer, transaction, etc. A digital twin can simulate one or more potential future states of a market, exchange, product, seller, buyer, transaction, etc. based on simulation parameters provided by the user. Examples of simulation parameters include progress over a period of time, potential actions by parties such as buyers or sellers, increases in supply and/or demand for products, resources, etc., changes in government regulations, and any other suitable parameters.

In embodiments, autonomous additive manufacturing platform 10110 may implement gamification in distributed manufacturing network 10130 by awarding points to various entities for performing tasks desirable to the operation of distributed manufacturing network 10130. For example, points may be awarded for trading parts or products of certain types and/or within certain areas. Entities awarded points may compete with one another, and digital and/or physical prizes may be awarded to entities that have achieved one or more point thresholds and/or ranked above one or more other entities on a points leaderboard.

In embodiments, scoring system 10634 may evaluate one or more manufacturing nodes or 3D printers within distributed manufacturing network 10130 based on customer satisfaction scores for meeting customer requirements. In embodiments, scores may form another basis for matching customers to manufacturing nodes or 3D printers.

In an embodiment, scoring system 10634 crowdsources customer satisfaction scores from multiple entities within distributed manufacturing network 10130. Examples of crowd sources include authenticating entities, domain experts, customers, manufacturers, wholesalers, and any other appropriate parties.

In embodiments, a certification entity or domain expert may authenticate one or more 3D printed parts as being of good quality, accurate, and/or reliable. In embodiments, a customer may review and certify one or more 3D printed parts or products, such as indicating that the part or product is in working order and/or of expected quality. In embodiments, manufacturers and/or wholesalers may sign instances of a 3D printed instruction set, such as applying a serial number to a 3D printed instruction set piece, before the 3D printed instruction set is available for transfer to a customer. Certifications, reviews, signatures, and/or any other verification indications made by a crowd source are recorded in the distributed ledger, such as by adding to the distributed ledger one or more new blocks indicating the verification, review, signature, or other verification indications. It can be.

In an embodiment, autonomous additive manufacturing platform 10110 operates with distributed manufacturing network 10130 to train models in artificial intelligence system 10212 to predict and manage product demand from one or more customers of distributed manufacturing network 10130. Use the system to learn based on a training set of results, parameters, and data collected from data sources associated with.

In embodiments, autonomous additive manufacturing platform 10110 may use results, parameters, and data collected from data sources associated with distributed manufacturing network 10130 to train models in artificial intelligence system 10212 to predict and manage material supply. It uses a system to learn based on a training set of data.

In an embodiment, autonomous additive manufacturing platform 10110 may use distributed manufacturing network 10130 to train models in artificial intelligence system 10212 to optimize production capacity for the distributed manufacturing network enabled by the autonomous additive manufacturing platform. Use the system to learn based on a training set of results, parameters, and data collected from data sources associated with.

In an embodiment, autonomous additive manufacturing platform 10110 uses the system to learn based on a training set of results, parameters, and data collected from data sources associated with distributed manufacturing network 10130 to form an artificial intelligence system 10212. Train a model to schedule across multiple production processes, printers, and manufacturing nodes, and dynamically recalibrate schedules based on changes in real-time production and priority data.

In an embodiment, autonomous additive manufacturing platform 10110 may be configured to manage a set of permission keys that provide access to one or more instances of a 3D printing instruction set 10902 and/or service associated with distributed manufacturing network 10130. You can use the ledger.

In an embodiment, the distributed ledger provides verifiable access to the 3D printing instruction set 10902, such as by one or more cryptographic authentication and/or techniques.

In embodiments, a distributed ledger may provide provable access to the 3D printing instruction set 10902 by one or more zero-knowledge proof techniques.

In embodiments, autonomous additive manufacturing platform 10110 may manage a distributed ledger to facilitate cooperation and/or cooperation between two or more entities regarding one or more instances of 3D printing instruction set 10902.

In embodiments, a trusted authorizing authority (e.g., autonomous additive manufacturing platform 10110 or another appropriate authority) may issue a private key and public key pair to each registered user of distributed manufacturing network 10130. Private and public key pairs can be used to encrypt and decrypt data (e.g., messages, files, documents, etc.) and/or perform operations on a distributed ledger.

In embodiments, autonomous additive manufacturing platform 10110 or other appropriate permissions may provide two or more levels of access to users.

In embodiments, autonomous additive manufacturing platform 10110 may define one or more classes of users, where each class of users is granted a respective access level.

In embodiments, autonomous additive manufacturing platform 10110 may issue one or more access keys to one or more classes of users, where each of the one or more access keys corresponds to a respective access level, thereby providing the user with their respective Provides different access levels through issued access keys.

In embodiments, possession of a particular access key may be used to determine the level of access to the distributed ledger. For example, a first class of users may be granted full observation access of a block, while a second class of users may be granted observation access of a block and the ability to verify and/or authenticate one or more instances of transactions contained within the block. Both may be granted, and a third class of users may have observation access to the block, the ability to verify and/or authenticate one or more instances of transactions contained within the block, and the ability to modify one or more instances of transactions contained within the block. Ability can be approved. In some embodiments, a class of users may be verified as legitimate users of the distributed ledger in one or more roles, and may grant relevant permissions to the distributed ledger and content stored therein.

In embodiments, distributed manufacturing network 10130 may establish a whitelist of trusted parties and/or devices, a blacklist of untrusted parties and/or devices, or a combination thereof to manage access.

In embodiments, additive manufacturing platform 10110 may be configured to create customized products for shoppers (i.e., customers) in or moving into a retail environment. Customized products may be printed in a retail environment by additive manufacturing unit 10102, thereby attracting customers to the retail environment. Customized products may include either or both decorative design and functional design. The decorative design may be configured to have one or more aesthetic elements that are customized according to the customer's profile. A functional design may be configured to have one or more functional features that are customized according to the customer's profile. For example, an additive manufacturing platform may use customer profile information, such as location data and/or search data, to determine that a customer will visit a retail environment. When a customer determines that they will visit a retail environment, the additive manufacturing platform can use information indicating the customer's aesthetic and/or functional desires to design a customized product for the customer. Additive manufacturing unit 10102 may manufacture customized products such that the customized products can be purchased by a customer from a retail environment. A customized product may be a product tailored to fit the customer's physiology. For example, a customized product may be the case for a cellular phone designed to fit a customer's hand based on data related to the shape and/or size of the customer's hand.

In embodiments, additive manufacturing platform 10110 may be configured to generate customized product samples to shoppers. Additive manufacturing platform 10110 may use data from customer profiles to determine one or more types of product samples that may appeal to the customer. Additive manufacturing unit 10102 may print product samples that appeal to customers prior to and/or during their visit to the retail environment. Product samples may include, for example, material samples, fabric samples, food samples, or any other suitable type of product sample.

In embodiments, additive manufacturing platform 10110 may be configured to use images, text, and/or video associated with a customer to build a customer profile. Images, text, and/or video may be sourced from one or more of web crawlers, social media feeds, public databases, etc.

In embodiments, additive manufacturing platform 10110 may include an AI system 10212 configured to perform AI and/or machine learning tasks related to the functionality of the additive manufacturing platform. AI system 10212 may be configured to design, at least in part, customized products for shoppers. AI system 10212 may use one or more machine learned models 10213 to analyze a customer profile and determine one or more customized products or features thereof that would be desirable to the customer. AI system 10212 may use one or more machine learning models 10213 to analyze sources of images, text, and/or video to build a customer profile. Machine learned model 10213 may be configured to allow AI system 10212 to determine types of images, text, and/or video that are somewhat valuable and/or effective for building a customer profile. AI system 10212 may use one or more machine learning models 10213 to determine the type of custom design that may be more or less desirable for the customer.

In embodiments, additive manufacturing platform 10110 may be configured to produce out-of-stock and/or low-stock products on-site in a retail environment. The platform may receive data related to the amount of inventory of products in the retail environment. The platform may determine that one or more products are out of stock and/or may be out of stock. AI system 10212 may be configured to determine which products are out of stock. When determining that one or more products are out of stock and/or may be out of stock, the platform may use the additive manufacturing unit 10102 to produce more products.

In embodiments, additive manufacturing platform 10110 may be configured to create infrastructure for a retail environment. The infrastructure may be a new infrastructure and/or a replacement infrastructure. The infrastructure may be created via additive manufacturing unit 10102. Examples of infrastructure include pallets, storage racks, display environments, signs, packages, tags, escalator components, elevator components, etc. Additive manufacturing platform 10110 can be configured to automatically determine the infrastructure needs of a retail environment. AI system 10212 may be configured to use machine learning models to determine and/or predict infrastructure needs of a retail environment.

In embodiments, an additive manufacturing platform may be configured to create customized products for shoppers (i.e., customers) in or moving into a retail environment. Customized products can be printed in a retail environment by 3D printing devices, thereby attracting customers to the retail environment. Customized products may include either or both decorative design and functional design. The decorative design may be configured to have one or more aesthetic elements that are customized according to the customer's profile. A functional design may be configured to have one or more functional features that are customized according to the customer's profile. For example, an additive manufacturing platform may use customer profile information, such as location data and/or search data, to determine that a customer will visit a retail environment. When a customer determines that they will visit a retail environment, the additive manufacturing platform can use information indicating the customer's aesthetic and/or functional desires to design a customized product for the customer. 3D printing devices can manufacture customized products so that the customized products can be purchased by customers from a retail environment. A customized product may be a product tailored to fit the customer's physiology. For example, a customized product may be the case for a cellular phone designed to fit a customer's hand based on data related to the shape and/or size of the customer's hand.

In embodiments, an additive manufacturing platform may be configured to generate customized product samples to shoppers. The additive manufacturing platform may use data from customer profiles to determine samples of one or more types of products that may appeal to the customer. A 3D printing device may print samples of products that appeal to customers before and/or during their visit to a retail environment. Product samples may include, for example, material samples, fabric samples, food samples, or any other suitable type of product sample.

In embodiments, an additive manufacturing platform may be configured to use images, text, audio, and/or video associated with a customer to build a customer profile. Images, text, audio, and/or video may be sourced from one or more of web crawlers, social media feeds, public databases, etc.

In embodiments, an additive manufacturing platform may include an AI system configured to perform AI and/or machine learning tasks related to the functionality of the additive manufacturing platform. The AI system may be configured to design, at least in part, customized products for shoppers. The AI system may use one or more machine learning models to analyze customer profiles and determine one or more customized products or features thereof that are desirable to the customer. The AI system may use one or more machine learning models to analyze the source of images, text, and/or video to build customer profiles. The machine learned model can be configured to allow the AI system to determine the types of images, text, and/or video that are somewhat valuable and/or effective for building customer profiles. The AI system may use one or more machine learning models to determine the type of custom design that may be more or less desirable for the customer.

In embodiments, an additive manufacturing platform may be configured to produce out-of-stock and/or low-stock products on-site in a retail environment. The platform may receive data related to the amount of inventory of products in the retail environment. The platform may determine that one or more products are out of stock and/or may be out of stock. The AI system can be configured to determine rewarehousing needs. Upon determining that one or more products are out of stock and/or may be out of stock, the platform may, by means of a 3D printing device, produce more products.

In embodiments, an additive manufacturing platform may be configured to create infrastructure for a retail environment. The infrastructure may be a new infrastructure and/or a replacement infrastructure. The infrastructure can be created through 3D printing devices. Examples of infrastructure include pallets, storage racks, display environments, signs, packages, tags, escalator components, elevator components, etc. The additive manufacturing platform can be configured to automatically determine the infrastructure needs of a retail environment. The AI system may be configured to use machine learning models to determine and/or predict the infrastructure needs of the retail environment.

In embodiments, an additive manufacturing platform 10110, e.g., associated with a value chain or other network, may be designed, prepared, configured, and/or designed to support the design, development, manufacture, and distribution of health and medical devices, components, parts, equipment, etc. can be placed. Facilitates the rapid delivery of medical and healthcare hard goods and devices produced by mobile additive manufacturing unit 10102 and/or additive manufacturing unit 10102, for example, in connection with a patient consultation with a medical or health care provider. Additive manufacturing units, such as units located in close enough proximity to a medical or health service provider, may be designated to assist with the consultation.

Based on the nature of the health care consultation (e.g., medical specialty and corresponding devices, equipment, and components), additive manufacturing unit 10102 can create a variety of possible health and medical devices to support health care providers and their patients. , may be equipped with suitable materials, such as metal and/or plastic printing materials, or combinations of other printing materials, suitable for printing components, parts, equipment, etc.

In embodiments, platform 10110 may include input from or related to a healthcare consultation, such as input indicating a needed medical device or part (e.g., technical specifications, CAD design, etc.); Input representing patient-specific data (e.g., clinical criteria, dimensions, measurements such as weight, height, girth, circumference, etc.); and input provided by medical and health care providers or other third parties, such as device specifications, requirements, etc. (e.g., limitations on device size such as thickness, requirements related to load- or stress-bearing minimums, or some other standards) can be taken.

In embodiments, platform 10110 may provide medical records (e.g., patient measurements, substance allergies, use of other related medical devices, etc.), device specification data (e.g., permissions for the device, part to be manufactured, or other object). manufacturing specifications from the party(s) holding the data), patient-input data (e.g., aesthetic preferences such as the color of the device), healthcare provider-input data (e.g., medical office branding), or some other input. It can process input from multiple sources, including but not limited to. In embodiments, an artificial intelligence system (e.g., a robotic process automation system trained on a training set of expert medical devices or other data) may generate recommended actions, prototypes, etc. that may involve the production of a device and/or components of a device. This is to determine the device. Additive manufacturing platform 10110, in some such embodiments, allows medical devices to be manufactured (e.g., using artificial intelligence systems, such as robotic process automation trained on expert data sets) from manufacturers (devices currently in stock and/or on order). ) and/or whether an additive manufacturing system should be used to produce the device, e.g., to meet immediate patient needs, reduce costs, etc. Similarly, additive manufacturing platforms may, in some embodiments, facilitate repairs, such as when complementary components can be created to replace worn or missing elements of a medical device using similar systems. It can be automatically determined that an element should be additively manufactured for this purpose.

In an example embodiment, an outpatient may visit an orthopedic office for a healthcare consultation related to a knee injury. Given the likelihood that a patient will require some form of external knee support from a medical device, such as a brace, prior to a health care consultation, the attending physician may determine whether the additive manufacturing platform 10110 will provide the necessary medical device. A user interface to the additive manufacturing platform, to determine the availability of knee braces and other medical devices desired to be manufactured by the additive manufacturing platform (to determine whether they have available designs, CAD renderings, and/or other specifications that enable production) , you can access the dashboard or some other user portal. If the additive manufacturing platform 10110 has these device specifications, the attending physician (or other staff associated with an upcoming patient healthcare consultation) can queue, reserve, or potentially interest the desired device design for manufacturing. Some other means of recording can be arranged. By having this record, at the time of the patient encounter, the attending physician (or other staff member associated with the patient's upcoming healthcare consultation) can use the user interface, dashboard, or some other user portal to the additive manufacturing platform to determine which device to select. Options may be presented to the patient. If the required medical device is not currently associated with an additive manufacturing platform, this may cause the platform to automatically send requests for corresponding device specifications, designs and other data needed to manufacture the device, component or part. Once these corresponding device specifications, designs, and other data are located, an alert will be sent to the attending physician (or other person involved in the upcoming patient healthcare consultation) indicating that there is a proposed product/device worthy of review that appears to comply with the enumerated device requirements. may be provided back to employees. As part of the review of each available specification, design or other data needed to manufacture the device, contract terms related to cost, warranty and other considerations may be presented for review. Contractual terms and contractual relationships between users of the additive manufacturing platform and third party holders of rights related to device manufacturing may be coordinated using smart contracts as described herein. Before, during, or after a patient's healthcare consultation, a medical device design may be selected and entered for manufacturing into an additive manufacturing platform. As part of the order, data related to a particular patient may be submitted to the additive manufacturing platform, such as data regarding the circumferences of the patient's lower leg, knee, and upper leg needed to create an appropriately sized brace. Such information may be entered into the additive manufacturing platform manually, or by transfer of data from a data source external to the additive manufacturing platform 10110, such as an electronic medical record, or some other data source that stores data related to device characteristics. It can be automatically entered into the additive manufacturing platform. Additional preferences, such as children who want an image of a koala bear engraved on the outside of their braces, or business people who want their braces to be a specific color to better match their skin tone and/or business suit color to make the braces less explicit. Data may also be provided. A user interface, dashboard or some other user portal for the additive manufacturing platform that allows a user, such as a patient, to view different prototypes and aesthetic decorations of the device to be manufactured prior to submitting work to be built. can enable interaction. Upon completing the design specifications, the additive manufacturing platform 10110 may proceed to produce the device and/or components or parts of the device while the patient's medical consultation is in progress, or such manufacturing may be completed after the consultation. , the device is automatically sent to the patient and/or healthcare provider based on contact data input entered into the additive manufacturing platform 10110 when placing an order.

In embodiments, for example, an additive manufacturing platform 10110 associated with a value chain network may be prepared, configured, and/or deployed to support printing of customized and/or personalized hotel textiles for a set of hotel guests. In one example, a unit located in sufficient proximity to a hotel to facilitate rapid delivery of mobile additive manufacturing unit 10102 and/or items produced by additive manufacturing unit 10102 in connection with an upcoming hotel guest visit. An additive manufacturing unit 10102 such as may be designated for support. In embodiments, textiles that can be customized and/or personalized may include bedding, sheets, towels, robes, pillows, blankets, curtains, furniture, etc.

In embodiments, additive manufacturing unit 10102 may be equipped with suitable materials, such as combinations of fabrics and other printing materials suitable for printing a variety of possible textiles, or other elements to support a hotel visit. In embodiments, the fabric may be canvas, cashmere, chenille, chiffon, cotton, crepe, damask, georgette, gingham, jersey, lace, leather, linen, merino wool, modal, muslin, organza, polyester, satin, silk, spandex. , suede, taffeta, twill, tweed, twill, velvet, viscose, etc., but is not limited thereto.

In an embodiment, the additive manufacturing platform 10110 may include input related to an upcoming hotel visit, such as an input indicating the type(s) of item(s) to be printed (e.g., pillows, bedding, towels, etc.); An input indicating the fabric type (such as cotton, silk, etc.); An input indicating the size of the item (such as to fit a queen bed or king bed); and input captured by cameras, microphones, data collectors, sensors, and other information sources associated with an upcoming hotel visit. For example, hotel staff can capture information related to hotel guest preferences. In an embodiment, the additive manufacturing platform 10110 may process the input, e.g., using an artificial intelligence system 10212 (such as a robotic process automation system trained on a training set of professional service visit data) to make recommendations. An action may be determined, which, in embodiments, may involve printing a fabric. In some such embodiments, the additive manufacturing platform 10110 may use an artificial intelligence system 10212 (e.g., robotic process automation trained on expert data sets) to determine whether the additive manufacturing unit 10102 should produce a textile. ) can be determined automatically.

In any such embodiment that recommends or configures instructions for additive manufacturing, the additive manufacturing platform 10110 discovers available materials/fabrics, configures instructions, initiates additive manufacturing, and determines when the elements are ready for use. You can provide hotel staff with updates, such as updates on what will happen.

In an embodiment, the additive manufacturing platform 10110 can determine, for example, through a trained AI agent, what other tasks are being done (e.g., to allow proper sequencing of the additive manufacturing output that aligns with the overall workflow). to a set of additive manufacturing units 10102 while recognizing the status of other relevant entities involved in other workflows, such as the priority of the printing job (e.g., whether it involves a loyal hotel guest), or other factors. You can automatically organize and schedule sets of tasks across tasks. In embodiments, optimization of workflow across a set of additive manufacturing entities may be achieved by having the artificial intelligence system 10212 perform a set of simulations, such as simulations involving alternative scheduling sequences, design configurations, alternative output types, etc. You can. In embodiments, simulations may be performed on additive manufacturing and other manufacturing entities (e.g., those performing cutting, dyeing, etc.), including handoffs between sets of different manufacturing entity types, such as when handoffs are handled by a robotic handling system. may include sequences involving cutting manufacturing entities and/or finishing entities sewing, constructing, adding hotel guest initials, etc.). In an embodiment, a set of digital twins may be comprised of the attributes and capabilities of various manufacturing systems, various handling systems (robotic systems, arms, conveyors, etc., but also human workforce), and/or the surrounding environment (e.g., hotels, manufacturing facilities, etc.). can indicate.

In embodiments, additive manufacturing platform 10110, e.g., associated with a value chain network, may be prepared, configured, and/or deployed to support restaurant operations. For example, in connection with a customer reservation at a restaurant, a table-side additive manufacturing unit 10102 and/or a portable unit such as a portable unit to facilitate direct delivery of items produced by the additive manufacturing unit 10102 to the table. Manufacturing unit 10102 may be designated to support customer reservations.

Based on the nature of the reservation (e.g., special dietary requirements, accessibility requirements, reservation opportunities) and the services and supplies available at the restaurant, the additive manufacturing unit 10102 may select various possible service items, special flatware, etc. , suitable materials such as food grade service/storage materials and combinations of other printing materials suitable for printing customized souvenir/feeding items, or other elements to support reservations. In an embodiment, the additive manufacturing platform 10110 may include inputs indicating, for example, time of day, size of party, special requests, partnership with restaurant owner, loyalty participation, etc.; input from or in connection with a reservation; Inputs indicating service support capabilities at the restaurant and options for timely access to locally available service support materials/equipment (such as status of ovens, countertops, food storage, meal preparation ingredients, customizable service items, etc.); and cameras, microphones, data collectors, sensors, and other information sources associated with the reservation, including optional input capture device(s) (e.g., a personal mobile phone with image capture features) associated with one or more participants in the reservation. You can take the input captured by For example, the host station camera may capture a set of photos of the scheduled participant(s), such as images of their faces, suitable for the creation of a 3D data set for additive manufacturing printing use.

In embodiments, additive manufacturing platform 10110 may process input, such as using artificial intelligence system 10212 to determine recommended actions for servicing participants in a reservation, which in embodiments may include: Standard service items adapted to meet the service requirements of the reservation, such as customized serving trays with separate compartments for each participant in the reservation, flatware and/or serving spoons adapted for human use without general attachments. It may involve the use of service items such as items of . Additive manufacturing platform 10110 may, in some such embodiments, use an artificial intelligence system 10212 trained on, e.g., expert data sets, etc. to determine whether a suitable service item is readily available and/or determine service item(s). can automatically determine whether using an additive manufacturing system to create can reduce delays, save costs, etc. Similarly, additive manufacturing platform 10110 may, in some embodiments, utilize a similar system to allow complementary components to be created to replace worn or missing components, such as a gas setting knob on a gas range regulator. Where possible, it may be automatically determined that elements should be additively manufactured to facilitate the use of additional kitchen equipment, such as countertops, to ensure timely meal service for reservations.

In embodiments, automated decisions may be made using a machine vision system that captures a set of facial images of reservation participants and generates a set of instructions for additive manufacturing a complementary service item, such as a glass cup, that matches the facial images. In any such embodiment recommending or configuring instructions for additive manufacturing, the additive manufacturing platform 10110 discovers available additive manufacturing units 10102 (e.g., a glass cup additive manufacturing unit in a restaurant building), You can configure compatible commands, initiate additive manufacturing, and provide updates to service staff, such as updates on when your custom printed glasses will be ready for use. In embodiments, the additive manufacturing platform 10110 may monitor the status of other related reservations in restaurants and other kitchen/service workflows, e.g., through trained AI agents (e.g., immediately at the start of the reservation). Timing of food preparation/meal courses (to allow for deprioritization of additive manufacturing operations that generate reservation-related service items that will not be used), additive manufacturing output aligned with overall kitchen workflow, meal service, etc. What other additive manufacturing operations are being done for other reservations (to allow proper sequencing of Additive manufacturing unit 10102 (glass cup additive manufacturing unit, kitchen equipment parts additive manufacturing unit, take-out/to-go food storage) or other factors (low service tip, free food/beverage items as compensation for delays, etc.) The system can automatically organize and schedule sets of tasks across sets (such as additive manufacturing units).

In embodiments, restaurant service items that can be enhanced and/or produced through additive manufacturing technologies include, but are not limited to, keeping a salad cool, keeping a hot meal warm, and keeping a serving of crispy French fries. , takeout/away containers configured to meet individual food item needs, containers shaped to meet food service item sizes/shapes (e.g., triangular-sized containers for sliced pies, round for pancakes, sandwich items) Includes rectangles/squares for . In embodiments, user-specific platformware, such as age range-specific platformware suitable for use by babies who have only learned to use a fork and spoon or children who are practicing their skills using a knife, (associate with reservation) These include non-traditional flatware items based on user preferences (expressed explicitly) or implicitly derived from user situations/images. Additionally, in embodiments, table and service items such as mugs, coasters, chargers, plates, etc. may be created to meet reservation aspects such as logos supplied with the reservation, opportunity specific designs/decor recommended during the reservation process, etc. In embodiments, an artificial intelligence system may be used to perform a set of simulations, such as simulations involving alternative food preparation and/or scheduling sequences, design configurations, alternative output/material types, etc., across a set of additive manufacturing entities/units. Optimization of workflow can be achieved.

In embodiments, a scheduled service item that relies on a blend of additive manufacturing materials, such as a paper-like material and an insulating structure (e.g., to maintain internal temperature and improve the comfort of the user holding the glass), may be used as a service item (e.g., to maintain internal temperature and improve user comfort while holding the glass). It can provide performance advantages over single material items, such as lower heat transfer from the interior of a custom printed glass cup (for example, a custom printed glass cup) to the exterior of the item.

In embodiments, additive manufacturing platform 10110, e.g., associated with a value chain network, may be prepared, configured, and/or deployed to support printing of personalized food products on university and/or corporate campuses. In one example, additive manufacturing unit 10102 may be designated to provide ethnic and personalized food to traveling students and workers. In embodiments, additive manufacturing unit 10102 may be equipped with materials suitable for printing a variety of possible food items to assist students or workers, such as combinations of ingredients and other printing materials. For example, pizza making can be automated by additive manufacturing unit 10102, where multi-nozzle print heads can top pizza with dough, sauce, and cheese along with a personalized selection of toppings. Similarly, desserts, chocolates, cakes, pastries, and even edible plates, utensils and cutlery, etc. can be printed by the additive manufacturing unit 10102.

In an embodiment, the additive manufacturing platform 10110 may include input from or associated with a customer, such as input indicating the type(s) of food item (e.g., pizza, pasta, dessert, etc.) to be printed; Input indicating taste preference (e.g. spicy, sweet, etc.); Input indicating aesthetic preferences (such as texture, color, etc.); An input indicating the food item size (e.g., small, medium, or large); inputs indicating nutritional requirements (e.g., proteins, carbohydrates, fats, vitamins, minerals, etc.), inputs indicating health needs (e.g., allergies, etc.), and cameras, microphones, data collectors, sensors, and other information associated with the upcoming campus visit. It can take input captured by the source, or some other input type. For example, information related to customer biological information may be captured to determine that the customer does not have any seafood allergies. In an embodiment, the additive manufacturing platform 10110 may process the input, e.g., using an artificial intelligence system 10212 (such as a robotic process automation system trained on a training set of professional service visit data) to make recommendations. Actions may be determined, and recommended actions may, in embodiments, involve, for example, printing customized sushi that optimizes ingredients to meet the customer's nutritional requirements.

In embodiments, additive manufacturing units 10102 may be used to keep salads cool, hot meals warm, maintain servings of crispy French fries, containers shaped to meet food service item sizes/shapes, etc. Takeout containers can be printed to meet the needs of individual food items such as:

In embodiments, food items may be printed in mobile additive manufacturing unit 10102 at or near the point of use on an on-demand basis, thereby reducing food inventory and costs associated with storage and transportation.

In an embodiment, the additive manufacturing platform 10110 may determine, for example, through a trained AI agent, what other tasks are being done (e.g., to allow proper sequencing of the additive manufacturing output that is aligned with the overall workflow). Additive manufacturing unit 10102 (e.g., based on the timing of customer orders) while being aware of the status of other related entities involved in other workflows, such as the priority of a printing job (e.g., based on the timing of a customer order), or other factors. It can automatically organize and schedule sets of tasks across sets (units that produce food, desserts, plates, cutlery, cutlery, kitchen equipment, etc.). In embodiments, optimization of workflow across a set of additive manufacturing entities may be achieved by having an artificial intelligence system perform a set of simulations, such as simulations involving alternative scheduling sequences, design configurations, alternative output types, etc. In embodiments, simulations may be performed on additive manufacturing and other manufacturing entities (e.g., those performing cutting, drilling, etc.), including handoffs between sets of different manufacturing entity types, e.g., when the handoffs are handled by a robotic handling system. May include sequences involving cutting manufacturing entities and/or finishing entities (including decorations, plates, garnishes, arrangements, glazes, etc.).

In embodiments, additive manufacturing platform 10110 may use live input, library data, personal data, licenses, etc. to autonomously design and create unique parts associated with a live event, such as personalized mementos, sample products, limited edition artwork, etc. It may be configured as a fixed or mobile system operating individually or as part of a network to compile data, etc.

In embodiments, additive manufacturing platform 10110 may use 3D scanning, such as a laser or white light scanner, image recognition, photography, publicly available data, etc. to obtain real-time or personalized input from a user or location, and store information. By combining and processing with existing public or licensed parts and data libraries, assembled 3D printable data sets and finished products can be created that can be waiting for a customer or later delivered to a home, business, or venue seat. You can.

In embodiments, additive manufacturing platform 10110, such as associated with a value chain network, may be configured and deployed by first responders to support first responder events. For example, in the context of a first responder request, additive manufacturing unit 10102 may be designated to support designing and printing custom components, parts, equipment, medical devices, accessories, etc. on an on-demand, real-time basis. Some examples of equipment that can be printed include personal protective equipment (PPE), face shields, goggles or medical glasses, protective eyewear, boots, surgical hoods, earplugs, valves, nozzles, helmets, body armor, extraction tools, etc. do.

In embodiments, equipment may be printed near or at the point of use as needed. For example, eyewear, earmuffs, helmets, and boots can be custom printed based on patient measurements. Similarly, equipment including respirators, ventilators, custom valves and nozzles can be printed on mobile additive manufacturing platforms based on immediate patient needs and delivered at the point of care.

In embodiments, the additive manufacturing platform 10110 may automatically determine (e.g., using an artificial intelligence system 10212 trained on an expert data set) that one or more parts should be additively manufactured to facilitate repair; , for example, where complementary parts may be created to replace worn or missing elements of first responder equipment or devices. Additive manufacturing platform 10110 may then process the input, such as by using artificial intelligence system 10212, to determine recommended actions for servicing the repair request.

In embodiments, a set of additive manufacturing units 10102 may be provided as a shared resource for multiple tenants of a building, such as a commercial real estate building, where the additive manufacturing units 10102 may provide networking resources (e.g., RF, Integrated with other building resources such as cellular, Wifi, optical fiber and other resources), computational resources (e.g., data storage resources, edge and cloud computing resources), IoT resources (e.g., cameras, sensors, etc.), The capabilities of the additive manufacturing unit 10102 may depend on the terms and conditions of the lease (which, in embodiments, may be embodied, at least in part, as a smart contract operating on data from or about the additive manufacturing unit 10102). It can be accessed by . In an embodiment, the additive manufacturing platform 10110 may be configured to (switches, access points, gateways, routers, wireless mesh network systems, satellite systems, Wifi systems, long-range RF systems (such as LORA), Zigbee, Bluetooth, and other wireless systems. 5G and other cellular network devices and infrastructure, as well as fixed networks, such as modems and other gateway devices for fiber access gateways and other systems, cable, Ethernet, digital subscriber lines, analog telephone lines, and other wired networking systems. Systems - each using any of a wide range of protocols such as Ethernet, TCP/IP, UDP, etc. - including devices, systems, services and systems within the backbone for buildings, campuses, etc., including a network backbone and/or a set of connectivity resources. Can contain, link to, or integrate with sets of other resources. Shared connection resources may include resources for Internet connections (e.g., wireless internet service provider (WISP) resources and fixed ISP connections), cellular connections (e.g., shared 5G), mesh network connections, etc. You can. In embodiments, the additive manufacturing platform 10110 may be a blockchain, distributed ledger, database or other data repository dedicated to a building, campus, etc., a distributed memory system that uses the memory of devices and systems that provide the building's IT infrastructure, etc. May contain, link to, or integrate with the same set of shared data storage resources. In embodiments, additive manufacturing units 10102 and other shared resources may be provisioned, e.g., by a host or a trained intelligent agent acting on behalf of a host, to provide (e.g., spare parts, products, tools, accessories, supplies, replacement parts, etc. may enable rapid customization and fulfillment of the needs of tenants, such as tenants of buildings, campuses, cities, etc., including operational requirements (such as for tenants of buildings, campuses, cities, etc.). Among many examples, additive manufacturing unit 10102 can be used to manufacture elements needed for specialty tenants such as personal protective equipment, respirators, wearable items, tools, etc., as well as IT-based components (such as connectors, plugs such as fiber optic cables, Ethernet ports, etc.) Elements necessary for structure can be created. In embodiments, shared resources may be monitored with various usage tracking technologies, such as event logs from networking nodes, logs from software systems, etc., assigning payment responsibilities, assigning usage rights (e.g., by tenant), , by time, by task, etc.) may be provisioned by an automated provisioning system that includes prioritizing resource utilization. This may include automated management by artificial intelligence agents trained by training data sets from expert resource managers. This can be a supervised, semi-supervised, or deep learning process and can include training on outcomes such as profitability outcomes, tenant feedback outcomes, user satisfaction outcomes, security outcomes, operational outcomes, etc. Resource sharing and payments may be governed and controlled by smart contracts, such as governing rules for allocating resources and conditional logic for prioritizing and/or determining payment responsibilities, and optionally on a distributed ledger of events involving resources. It works. In embodiments, the smart contract framework itself allows tenants to, e.g., provide services, share resources (e.g., with other tenants, including any of the resources mentioned in this application as well as other resources), etc. To enable this, it may be a shared resource provided to the tenant.

LIQUID LENS

122-127 relate to various embodiments and applications of liquid lens devices. Liquid lens devices can be used in a suite of applications, including autonomous systems that rely on image classification to perform tasks. Liquid lens devices can be integrated into many different areas of the value chain to improve the performance of various autonomous systems, especially by providing improved image sensing capabilities and image classification.

122 is a schematic diagram illustrating an example implementation of a conventional computer vision system 11100 for recognizing an object of interest 11102. Computer vision system 11100 includes a lens assembly 11104 that attempts to focus light from object 11102 onto sensor 11106. Sensor 11106 may be an image sensor, such as a charge coupled device (CCD) or complementary metal oxide semiconductor (CMOS) device that includes an array of photosensitive elements. The sensor can convert light into an analog electrical signal corresponding to light intensity. An analog-to-digital (AD) converter 11108 then converts the analog voltage to digital data. This raw digital data is then transmitted to image processing system 11110 for analysis. Image processing system 11110 then processes the raw digital data to produce image 11112. Image processing system 11110 may also involve pre-processing and post-processing, including image scaling, noise reduction, color adjustment, brightness adjustment, white balance adjustment, sharpness, adjustment, contrast adjustment, etc., to improve image quality. Additionally, images may be analyzed using machine learning or other algorithms to identify one or more objects within the image.

Conventional computer vision systems 11100 have many limitations. Attempting to recreate vision by creating a focused image results in the loss of a large amount of information, leaving the vision system 11100 with limited data. Computer vision systems 11100 typically generate two-dimensional images of three-dimensional objects and are unable to capture information related to aspects such as object depth, motion, orientation, etc. Algorithms in computer vision system 11100 limit the quality of inference by attempting to infer information about 3D scenes/objects from 2D frames and information.

FIG. 123 is a schematic diagram illustrating an example implementation of a dynamic vision system 11200 for dynamically learning object concepts regarding an object of interest 11202 according to an embodiment of the present disclosure. The dynamic vision system 11200 may replace and/or augment the lens 11104 of the conventional vision system 11100 with a variable focus liquid lens 11204. Variable focus liquid lens 11204 can be an electrically controlled cell containing an optical grade liquid that is deformed through an electric current to change the shape of the lens. The dynamic vision system 11200 measures focal length, spherical aberration, field curvature, coma, chromatics aberrations, distortion, vignetting, ghosting and flaring, and diffraction of light. This flexibility of the liquid lens 11204 is exploited by continuously adjusting the lens parameters to dynamically change the various optical properties of the light passing through the containing lens. Accordingly, a fully variable liquid lens allows for a more dynamic input to the sensor 11206, allowing the capture of visual information and metadata that would otherwise be lost in a conventional computer vision system 11100.

An analog-to-digital (AD) converter 11208 may generate digital data from the rich visual information captured by the sensor 11206 and an image processing system 11208 with pre-processing capabilities and the ability to post-process additional optical parameters as part of the image. You can create an image based on . Processing system 11209 may also include a control system 11212 configured to adjust in real time one or more optical parameters including focal length, liquid material, reflectivity, color, environment, and lens shape. Adaptive intelligence system 11214 may then dynamically learn based on the training set of results, parameters, and data collected from liquid lens 11204 to generate object concepts 11216. Object concept 11216 may include contextual intelligence about the object and the environment, which may then be processed by adaptive intelligence system 11214 to recognize object 11202.

In embodiments, adaptive intelligence system 11214 may process input data from the liquid lens and dynamically learn object concepts to provide superior object recognition and vision, such as using machine learning or other algorithms, neural networks, expert experts, etc. May include artificial intelligence capabilities, involving systems, models, etc.

In an embodiment, the adaptive intelligence system 11214 is an intelligence layer that receives requests from a set of intelligence layer clients and responds to these requests by providing intelligent services (e.g., decisions, classifications, predictions, etc.) 140).

In an embodiment, dynamic vision system 11200 may be used to capture a large amount of information that would otherwise be lost when inferring depth and distance from a focused image of a conventional vision system 11100. A real-time adjustable data stream may be supplied to processing system 11209 to generate situational awareness or generate out-of-focus images. Dynamic input to liquid lens 11204 can provide richer metadata for image processing because the image is based on additional optical parameters rather than just focal length and aperture. Image processing system 11210 may include information that was previously lost to generate a new set of insights about objects and their surroundings that were not captured by conventional computer vision systems 11100.

Compared to a conventional computer vision system 11100 that utilizes fixed sensory elements, the dynamic vision system 11200 provided in this application may utilize a dynamically learned liquid lens assembly. The compliant liquid lens 11204 within the assembly may be continuously and/or frequently adjusted, for example, based on environmental factors and/or feedback from the processing system 11209 to better understand the physical properties the image represents while providing deeper context. Training data corresponding to light can be generated. By training dynamic vision system 11200 to recognize objects using variable optical parameters through a liquid lens assembly, processing system 11209 can learn the optimal optical setting(s) for detecting the object. More dynamic input to the dynamic vision system 11200 may result in creating richer context and providing superior object recognition.

Dynamic vision system 11200 may integrate sensing, control, and processing functions and dynamically adjust liquid lens 11204 as vision algorithms in processing system 11209 take different inputs to produce real-world vision results. Adjust.

Dynamic vision system 11200 mimics biological vision by integrating sensing, control, and processing functions (biological vision involves a stream of information directly conveyed by deep learning systems, where these deep learning systems determine orientation, centration, and can directly alter aspects of vision processing, including attention, eyelid action, blinking, and communication with other humans).

In embodiments, dynamic vision system 11200 may use saccades to build rich models of objects in their environment by characterizing them by context and capturing situational intelligence through associations. This reflects how saccades capture information about objects in the environment. A saccade refers to the rapid, simultaneous movement of both eyes between two or more areas of focus. While observing a scene, the human eye makes several sporadic saccade stops, moving quickly between each stop, locating key parts of the scene, building a mental three-dimensional map corresponding to the scene. The dynamic vision system 11200 and methods described in this application may use saccades to characterize objects by context and allow control of the optical system to more quickly identify and characterize the field of view. Saccades integrate a variety of physical/optical properties with object-oriented learning, rapidly improving comprehension and retrieval in the visual sphere.

In embodiments, dynamic vision system 11200 may also mimic a human baby's biofeedback loop to create a system of associative memory and vision and build a causal three-dimensional model of the environment. The human baby's learning system involves many feedback loops of activity in which the baby builds a causal model of the world around him by performing a sequence of controlled experiments. The dynamic vision provided by liquid lens-based vision systems can mirror a baby's learning algorithm, in part, by starting a training set around an object and letting the learning algorithm figure out the right way to look at the object.

124 displays a schematic diagram illustrating an example architecture of a dynamic vision system 11300 showing details of various components in accordance with some embodiments of the present disclosure. A dynamic vision system 11300 for recognizing an object 11302 may include an optical assembly 11304 and a processing system 11306. Optical assembly 11304 may include a compliant liquid lens 11308, a sensor 11310, and an analog-to-digital (AD) converter 11312. The processing system 11306 may include a control system 11314, an image processing system 11316, an adaptive intelligence system 11318, a digital twin system 11320, and a simulation system 11322. Adaptive intelligence systems may include machine learning systems 11324 and artificial intelligence systems 11326.

The compliant liquid lens 11308 of the optical assembly 11304 is a sensor 11310 that is then provided to the processing system 11306 to generate situational awareness or computerized understanding of the world in which the dynamic vision system 11300 is operating. ) may be adjusted frequently in real time based in part on changes in one or more optical parameters by the control system 11314, which generates a real-time data stream. This understanding can include rich contextual intelligence about objects and the environment and can be expressed as object concepts. The object concept can be used by processing systems for object recognition, predicting object motion, position and orientation, creating 3D models of objects, monitoring objects for random defects, and other applications. For example, adaptive intelligence system 11318 may process object concepts to build a three-dimensional representation of the object. Machine learning system 11322 in adaptive intelligence system 11318 may input object concepts to one or more machine learning models, and the object concepts are used as training data for the machine learning models. Artificial intelligence system 11326 may also be configured to classify, predict, and make other decisions related to objects, including determining the location, orientation, and motion of the object.

In embodiments, dynamic vision system 11300 may be configured to process sensor information to create a three-dimensional representation of object 11302 in a single step without the intermediate step of processing it into a flat image.

In embodiments, control system 11314 may provide control commands to one or more actuators that in turn drive adjustment of the liquid lens configuration. An actuator may be actuated by an energy source, typically electric current, hydraulic fluid pressure, or pneumatic pressure, and may convert such energy into motion. Examples of actuators include linear actuators, solenoids, comb drives, digital micromirror devices, electric motors, electroactive polymers, hydraulic cylinders, piezoelectric actuators, pneumatic actuators, servomechanisms, servo motors, thermal bimorphs, screw jacks, or any other type of hydraulic, pneumatic, electrical, mechanical, thermal, magnetic type actuator, or some other type of actuator.

In embodiments, control system 11314 may provide control instructions to one or more actuators to change the focal length of the liquid lens based on a stimulus. This can provide autofocus capabilities to the dynamic vision system 11300 by focusing, refocusing, or defocusing the lens to a desired focal length. Stimulation mechanisms may include electrical, hydraulic, pneumatic, mechanical, thermal, or magnetic.

Some examples of control systems 11314 include electrowetting, sound piezoelectric, and electro-active polymers.

In embodiments, a compliant liquid lens assembly within the dynamic vision system 11300 may have an electrowetting control system such that application of a voltage to the fluid within the liquid lens changes the shape of the liquid, effectively changing the focus of the liquid lens assembly. there is.

In embodiments, the placement of actuators in a variable focused liquid lens based optical assembly may be optimized using machine learning.

In embodiments, control system 11314 may control liquid lens 11304 configuration based on feedback from processing system 11306 in response to changes in environmental factors. Some examples of environmental factors include temperature, vibration, ambient sensor data, workflow, entity ID, user behavior data, entity profiling, similarity to known data, etc.

In embodiments, control system 11314 responds to changes in source illumination, including controlling color, color temperature, timing (PWM), amplitude (e.g., increasing PWM but decreasing amplitude, direction, polarization, etc.). Thus, the configuration of the liquid lens 11304 can be controlled based on feedback from the processing system 11306.

In embodiments, control system 11314 may control liquid lens configuration based on human occupancy and recognition of when illumination needs to be coordinated with human needs versus when it needs to be coordinated only to serve the liquid lens system. .

In an embodiment, optical assembly 11304 may include multiple sets of liquid lenses with a processing system 11306 that coordinates control of multiple liquid lens setups.

In an embodiment, optical assembly 11304 may include multiple sets of liquid lenses, where each lens has a separate objective function, and separate processing systems with AI setups or algorithms.

In embodiments, optical assembly 11304 may include one or more liquid lenses in combination with conventional convex or concave optical lenses, and processing system 11306 coordinates control of the combination.

In embodiments, processing system 11306, e.g., using adaptive intelligence system 11318, digital twin system 11320, and simulation system 11322, can control mechanical, optical, or lighting aspects of dynamic vision system 11300. You can run simulations to model, simulate, and characterize. Simulations executed by processing system 11306 can help identify suitable imaging components for dynamic vision system 11300, including sensors, lenses, and lights. Simulations may include real-time analysis to calculate a wide range of metrics, build charts, graphs, and models, and visualize the impact of changes in one or more optical parameters on the performance of dynamic vision system 11300. Artificial intelligence system 11326 within adaptive intelligence system 11318 then uses one or more models to classify, predict, recommend, and/or determine the lens material, geometry, and optical properties of dynamic vision system 11300. , can generate or facilitate decisions or instructions related to performance and design. For example, artificial intelligence system 11326 may run a simulation on one or more liquid lens digital twins to generate recommendations regarding fluids used in the liquid lenses. The simulation uses different fluids including distilled water, methyl alcohol, ethyl alcohol, ether, carbon tetrachloride, methyl acetate, glycerin, nitrobenzene, etc. to generate recommendations for the preferred fluid for a given application of the dynamic vision system 11300. It can be done.

The dynamic vision system 11300 can utilize dynamically learned sensory elements to recognize objects, ensuring richer object recognition capabilities that can be applied to a very wide range of use cases. This approach is ideal for imaging applications requiring rapid focusing, high throughput, and control over depth of field and working distance. Moreover, this approach is particularly beneficial for complex vision applications where conventional vision techniques are inadequate. Some examples of these applications include: recognizing objects in dynamic environments, such as when the object or vision system is moving; recognizing three-dimensional (3D) objects by capturing depth data; recognizing small objects; recognizing facial features; recognizing objects in power-constrained or network-constrained environments; etc.

In embodiments, dynamic vision system 11300 is within or with a set of Value Chain Network (VCN) entities (such term includes many examples and embodiments disclosed in this application and documents incorporated herein by reference). can be integrated.

In embodiments, dynamic vision system 11300 may be integrated within or with a set of robotic systems, such as mobile and/or autonomous robotic systems. For example, dynamic vision system 11300 may be included within the housing or body of a robotic system, such as a multipurpose/general purpose robotic system that simulates human or other animal species capabilities. Vision capabilities may enable robots to identify and manipulate target objects for use in robotic assembly lines where object depth, orientation, position, and motion can be inferred for improved object identification. Vision capabilities can also enable a robot to simultaneously localize and map, a technique for estimating a robot's position relative to its surroundings while simultaneously mapping the environment. As another example, dynamic vision system 11300 may be integrated with a robotic exoskeleton designed to augment the capabilities of a human operator and provide optimized sensing and control for the human operator.

In an embodiment, the output from the dynamic vision system 11300 is temporally combined with output from other sensors within the robot using conditional probabilities to provide a richer representation of the object, including information about the object's position, orientation, and motion. You can create combined views. Some examples of sensors that may be used with the liquid lens based dynamic vision system 11300 include cameras, LIDAR, RADAR, SONAR, thermal imaging sensors, hyperspectral imaging sensors, light sensors, force sensors, torque sensors, speed sensors, and acceleration sensors. , position sensor, proximity sensor, gyro sensor, sound sensor, motion sensor, position sensor, load sensor, temperature sensor, touch sensor, depth sensor, ultrasonic range sensor, infrared sensor, chemical sensor, magnetic sensor, inertial sensor, gas sensor, Includes a humidity sensor, pressure sensor, viscosity sensor, flow sensor, object sensor, tactile sensor, or some other type of sensor.

In embodiments, a dynamic vision system 11300 comprising compliant liquid lenses controlled by AI and augmented by sensors as needed may be adapted to build a neuroprosthetic system.

In embodiments, a dynamic vision system 11300 including adaptive liquid lens technology controlled by AI as needed may be adapted to build an exoskeleton system.

In embodiments, a dynamic vision system 11300 comprising compliant liquid lenses controlled by AI and augmented by sensors as needed may be adapted to perform facial recognition on human faces obscured by facial masks.

Figure 125 shows a flowchart illustrating a method for object recognition by a liquid lens-based dynamic vision system according to some embodiments of the present disclosure.

Referring to Figure 125, at 11402, a real-time data stream representing object concepts is received from a liquid lens based optical assembly. The data stream may be received from the sensor and may contain rich context and visual information generated by continuously adjusting the liquid lens in response to changes in optical parameters. Data streams can be analyzed on edge devices or sent for data processing by local or remote intelligence. The use of cloud-connectable edge devices, e.g., within a computing infrastructure proximate to and/or integrated with or included in dynamic vision system 11300, may include, e.g. If the dynamic vision system 11300 has onboard edge computation and/or connectivity resources, such as 5G (or other cellular), Wi-Fi, Bluetooth, fixed networking resources, etc., it may provide rapid, real-time or near-real-time processing responsiveness. It can provide an opportunity. At 11404, the real-time data stream is processed by an image processing system to determine object concepts, including contextual intelligence about the object and its environment. At 11406, optical parameters are adjusted by the control system to induce changes in the configuration of the liquid lens. A continuously adjusting liquid lens generates a real-time data stream from the sensor and rich metadata for image processing because the image is based on additional optical parameters rather than just focal length and aperture. At 11408, the object concept is sequentially modified and used as input to train a machine learning model that dynamically learns based on a training set of results, parameters and data collected from the liquid lens-based optical assembly. At 11410, object concepts, including contextual intelligence about the object and the environment, are utilized by artificial intelligence to make classification, prediction, and other decisions related to the object, including determining the location, orientation, and motion of the object.

Figure 126 shows a schematic diagram illustrating an example implementation of a dynamic vision system for modeling, simulating, and optimizing various optical, mechanical, design, and lighting parameters of a dynamic vision system in accordance with some embodiments of the present disclosure. Dynamic vision systems can learn based on data captured by sensors in response to sequential adjustments of liquid lenses to train an artificial learning system to use the digital twin for classification, prediction, and decision making.

Digital twin system 11320 may be configured to simulate the behavior of dynamic vision system 11300 to continuously capture key operational metrics, and monitor and optimize performance of dynamic vision system 11300 in real time or near real time. It can be used to: Digital twin system 11320 may create a digital replica, or digital twin 502, of one or more of the components or subsystems of dynamic vision system 11300. The digital twin 502 of one or more of the components or subsystems may use real-time sensor data to provide a realistic, real-time virtual representation and simulation of one or more possible future states of the one or more components and subsystems. there is. Digital twin 502 may be continuously updated based on sensor data to reflect current conditions or parameter values of the component or subsystem. Therefore, a digital twin provides a high-fidelity digital simulation of the behavior of a component or subsystem. This capability can be used to create digital profiles of both the previous and current behavior of a component or subsystem, with the resulting profiles used to predict future behavior of the component or subsystem as well as detect suboptimal behavior. do.

126, the digital twin 11502 in the dynamic vision system 11300 includes an object twin 11504, an environment twin ( 11506), liquid lens twin (11508), optical lens twin (11510), sensor twin (11512), process twin (11514), actuator twin (11516), object concept twin (11518), etc. Digital twin 11502 may be populated with related data, for example, liquid lens twin 11508 may have a corresponding liquid lens twin including dimensional data, material data, geometry data, feature data, thermal data, vibration data, etc. It can be filled with relevant data. A digital twin may, in embodiments, provide one or more simulations of both the physical elements and characteristics of one or more replicated components or subsystems and their dynamics throughout the lifecycle of the one or more replicated components.

In embodiments, digital twin 11502 may be used during any suitable hypothetical situation, such as, for example, during a design phase before one or more components are manufactured or fabricated, or during or after construction or fabrication of one or more components, It can provide a virtual simulation of one or more components or subsystems by allowing virtual extrapolation of sensor data to simulate the state of one or more components. In an embodiment, machine learning model 11520 may, for example, predict possible improvements to one or more components, predict whether one or more components are compatible with each other, predict when one or more components may fail, etc. , and/or by suggesting possible improvements to one or more components, such as changes to parameters, arrangement, configuration, or any other suitable changes to the component. It can be predicted automatically. For example, liquid lens twin 11506 and optical lens twin 11510 can run a virtual simulation to check compatibility with each other as well as with the optical assembly and predict optimal arrangement within the assembly.

In embodiments, machine learning model 11520 along with digital twin 11502 can help drive a variety of applications including material selection 11522, design optimization 11524, and motion prediction 11526.

In embodiments, digital twin 11502 can be used during both the design and operation phases of one or more components by facilitating the observation, measurement, and analysis of various metrics and then conveying insights to the design or operation process for improvement of these processes. In addition to simulation of one or more components, it may also allow simulation of virtual operating conditions and configurations of one or more components.

Simulation system 11322 may set up, provision, configure, and otherwise manage simulations and interactions between digital twins 11502. Accordingly, the simulation system can help simulate, evaluate, and optimize the behavior and characteristics of the various components and subsystems of the dynamic vision system 11300 using the digital twin 11502 of these components and subsystems.

In embodiments, artificial intelligence system 11326 may be configured to run simulations in simulation system 11322 using liquid lens twin 11508 and/or other digital twins 11502 available to digital twin system 214. You can. For example, processing system 11306 may adjust one or more optical parameters of liquid lens twin 11508. In an embodiment, artificial intelligence system 11326 may, for each set of parameters, run a simulation based on the set of parameters and collect simulation result data resulting from the simulation. For example, the artificial intelligence system 11326 may run a simulation by varying the optical parameters of the liquid lens twin 11506 to generate simulation results in the form of an object concept twin 11518. During simulation, processing system 11306 may vary the focal length, fluid material, reflectivity, color, environment, lens shape, and any other parameters of liquid lens twin 11506. The resulting data from this simulation in the form of an object concept twin 11518, in addition to other sensor data as well as data from other sources, can then be used by the machine learning system 11324 to train a machine learning model 11520. .

In embodiments, machine learning model 11520 may process data received from sensors, including event data and state data, to define simulation data for use by digital twin system 11320. Machine learning model 11520 may, for example, receive state data and event data related to specific components of dynamic vision system 11300 and format the state data and event data in a format suitable for use by digital twin system 11320. A series of operations can be performed on status data and event data to format them. For example, machine learning model 11520 may collect data from one or more sensors located on, near, within, and/or around the liquid lens, process the sensor data into simulation data, and convert the simulation data into a digital twin. It can be output to the system 11320. Digital twin system 11320 may then use the simulation data to create a liquid lens twin 11506, where the simulation may include, for example, shape, material, focal length, reflectivity, environment, lighting, color, temperature, pressure, Includes metrics including wear and vibration. The simulation may be a substantially real-time simulation that allows a user of dynamic vision system 11300 to view a simulation of the liquid lens, its associated metrics, and the portion-related metrics in substantially real time. The simulation may be a predictive or virtual situation that allows a user of dynamic vision system 11300 to view a predictive or virtual simulation of a liquid lens, metrics associated therewith, and metrics associated with the component.

In embodiments, machine learning model 11520 and digital twin system 11320 process sensor data and create a digital twin for a set of components to perform real-time simulation, predictive simulation, and/or virtual simulation of related groups of components. It can be done easily.

The machine learning model 11520 may be an algorithm and/or statistical model that does not use explicit instructions but instead relies on patterns and inferences to perform a specific task. Machine learning model 11520 may build one or more mathematical models based on training data to make predictions and/or decisions without being explicitly programmed to perform a specific task. In example implementations, machine learning models may perform classification, regression, clustering, anomaly detection, recommendation generation, digital twin creation, and/or other tasks.

In embodiments, machine learning model 11520 may perform various types of classification based on input data. Classification is a predictive modeling problem in which class labels are predicted for given examples of input data. For example, machine learning model 11520 may perform binary classification, multi-class, or multi-label classification. In embodiments, a machine learning model may output a “confidence score” indicating the respective confidence associated with the classification of the input into each class. In embodiments, confidence scores may be compared to one or more thresholds to render discrete category predictions. In an embodiment, a certain number of classes (e.g., one) with the relatively greatest confidence score may be selected to render a discrete category prediction.

In embodiments, machine learning model 11520 may output a probabilistic classification. For example, machine learning model 11520 can predict a probability distribution for a set of classes, given sample input. Therefore, rather than outputting only the most likely class to which the sample input should belong, machine learning model 11520 may output, for each class, the probability that the sample input belongs to that class. In an embodiment, the probability distribution for all possible classes may sum to 1. In embodiments, a softmax function, or other type of function or hierarchy, may be used to convert a set of real values each associated with a possible class into a set of real values in the range (0, 1) that sums to 1. In embodiments, the probabilities provided by the probability distribution may be compared to one or more thresholds to render discrete category predictions. In embodiments, only a certain number of classes (e.g., one) with the relatively greatest prediction probability may be selected to render discrete category predictions.

In embodiments, machine learning model 11520 may perform regression modeling and related processes to provide output data in the form of continuous numeric values. By way of example, machine learning model 11520 may perform linear regression, polynomial regression, logistic regression, non-linear regression, or some other modeling process. As described, in embodiments, the softmax function or other function or layer squashes the set of real values each associated with two or more possible classes into a set of real values in the range (0, 1) that sums to 1. can be used to For example, machine learning model 11520 may perform linear regression, polynomial regression, or non-linear regression. As an example, machine learning model 11520 may perform simple regression or multiple regression. As described above, in some implementations, the softmax function or other functions or hierarchies convert a set of real values, each associated with two or more possible classes, into a set of real values in the range (0, 1) whose sum is 1. Can be used for quashing.

In embodiments, machine learning model 11520 may perform various types of clustering. For example, machine learning model 11520 may identify one or more previously defined clusters to which the input data most likely corresponds. In some implementations in which machine learning model 11520 performs clustering, machine learning model 11520 may be trained using unsupervised learning techniques.

In embodiments, machine learning model 11520 may perform anomaly detection or outlier detection. For example, machine learning model 11520 may identify input data that does not conform to expected patterns or other characteristics (e.g., as previously observed from previous input data). As an example, anomaly detection may be used to detect fraudulent transactions or detect system failures.

In some implementations, machine learning model 11520 may provide output data in the form of one or more recommendations. For example, machine learning model 11520 may be included in a recommender system or engine. As an example, given input data describing previous results (e.g., scores, rankings, or grades representing amounts of success or enjoyment) for a particular entity, machine learning model 11520 may: , can output suggestions or recommendations of one or more additional entities that are expected to have the desired results.

As described above, machine learning model 11520 may be or include one or more of a variety of different types of machine learning models. Examples of these different types of machine learning models are provided below for illustration. One or more of the example models described below may be used (e.g., combined) to provide output data in response to input data. Additional models other than the example models provided in this application may also be used.

In some implementations, machine learning model 11520 can be, for example, a linear classification model; secondary classification model; It may be or include one or more classifier models, such as the like. Machine learning model 11520 may include, for example, a simple linear regression model; multiple linear regression model; Logistic regression model; stepwise regression model; Multivariate adaptive regression spline; locally estimated scatterplot smoothing model; It may be or include one or more regression models, such as the like.

In some examples, machine learning model 11520 may include, for example, a classification and/or regression tree; chi-squared automatic interaction detection decision tree; decision stump; conditional decision tree; It may be or include one or more decision tree-based models, such as the like.

Machine learning model 11520 may be or include one or more kernel machines. In some implementations, machine learning model 11520 may be or include one or more support vector machines. Machine learning model 11520 may include, for example, a learning vector quantization model; Self-organizing map model; Locally weighted learning model; It may be or include one or more instance-based learning models, such as the like. In some implementations, machine learning model 11520 can be, for example, a k-nearest neighbor classification model; k-nearest neighbor regression model; It may be or include one or more nearest neighbor models, such as the like. Machine learning model 11520 may include, for example, a Naive Bayes model; Gaussian naive Bayes model; multinomial naive Bayes model; averaged 1-dependency estimator; Bayesian network; Bayesian trust network; Hidden Markov Model; It may be or include one or more Bayesian models, such as the like.

In some implementations, machine learning model 11520 may be or include one or more artificial neural networks (also referred to simply as neural networks). A neural network may contain groups of connected nodes, which may also be referred to as neurons or perceptrons. Neural networks can be organized into one or more layers. A neural network containing multiple layers may be referred to as a “deep” network. A deep network may include an input layer, an output layer, and one or more hidden layers located between the input layer and the output layer. Nodes in a neural network can be connected or completely disconnected.

Machine learning model 11520 may be or include one or more feedforward neural networks. In a feedforward network, the connections between nodes do not form cycles. For example, each connection may connect a node from a previous layer to a node from a later layer.

In some cases, machine learning model 11520 may be or include one or more recurrent neural networks. In some cases, at least some of the nodes of a recurrent neural network may form a cycle. Recurrent neural networks can be particularly useful for processing input data that is sequential in nature. In particular, in some cases, a recurrent neural network may transfer or retain information from a previous portion of an input data sequence to a subsequent portion of the input data sequence through the use of recursive or directed recurrent node connections.

In some examples, sequential input data may include time series data (e.g., time or images captured at different times subject to sensor data). For example, a recurrent neural network can analyze sensor data versus time to detect or predict swipe direction, perform handwriting recognition, etc. Sequential input data includes words within sentences (e.g., for natural language processing, conversation detection or processing, etc.); notes in musical composition; sequential actions taken by the user (e.g., to detect or predict sequential application usage); sequential object states; It may include etc.

Exemplary recurrent neural networks include long short-term (LSTM) recurrent neural networks; gated circulation unit; Bidirectional recurrent neural network; Continuous-time recurrent neural network; neural history compressor; echo state network; Elman Network; Jordan Network; recursive neural network; Hopfield Network; fully recurrent network; Includes sequence-to-sequence configuration, etc.

In some examples, machine learning model 11520 may be or include one or more acyclic sequence-to-sequence models based on self-attention, such as a transformer neural network.

In some implementations, machine learning model 11520 may be or include one or more convolutional neural networks. In some cases, a convolutional neural network may include one or more convolutional layers that perform convolutions on input data using learned filters.

A filter may also be referred to as a kernel. Convolutional neural networks can be particularly useful for vision problems, such as when the input data includes images such as still images or video. However, convolutional neural networks can also be applied for natural language processing.

In some examples, machine learning model 11520 may be or include one or more generative networks, such as, for example, an adversarial generative neural network. Generative networks can be used to generate new data, such as new images or other content.

Machine learning model 11520 may be or include an autoencoder. In some cases, the goal of an autoencoder may be to learn a representation for a set of data (e.g., low-dimensional encoding), typically for the purpose of dimensionality reduction. For example, in some cases, an autoencoder may attempt to encode input data and provide output data that reconstructs the input data from the encoding. Recently, the autoencoder concept has become more widely used to learn generative models of data. In some cases, autoencoders may involve additional losses beyond reconstructing the input data.

Machine learning model 11520 may be, for example, a deep Boltzmann machine; deep trust network; Stacked autoencoder; It may be or include one or more different types of artificial neural networks, such as the like. Any of the neural networks described in this application may be combined (e.g., stacked) to form more complex networks.

Machine learning model 11520 may include one or more clustering models, such as, for example, a k-means clustering model; k-median clustering model; expectation maximization model; hierarchical clustering model; It may include etc.

In some implementations, machine learning model 11520 may include, for example, principal component analysis; Kernel principal component analysis; Graph-based kernel principal component analysis; principal component regression; Partial least squares regression; Salmon Mapping; multidimensional scaling; projection tracking; linear discriminant analysis; mixed discriminant analysis; second-order discriminant analysis; generalized discriminant analysis; flexible discriminant analysis; auto-encoding; One or more dimensionality reduction techniques may be performed, such as:

In some implementations, machine learning model 11520 includes one or more reinforcement learning techniques, such as a Markov decision process; dynamic programming; Q function or Q-learning; value function approach; deep Q-network; differentiable neural computer; A3C (asynchronous advantage actor-critics); deterministic policy gradient; etc. can be performed or applied to it.

In an embodiment, the data processing system is implemented using a neural network to provide real-time adaptive control of the dynamic vision system 11300, including object classification and determination of object location, orientation, and motion.

In some embodiments, a neural network model may be used directly to determine adjustments to optical parameters using training or learning of the neural network model. Initially, the model may be allowed to randomly select from a range of values for each input optical control parameter or action. If an optical control parameter adjustment or sequence of actions leads to incorrect prediction/classification, it can be scored as leading to an undesirable (or negative) outcome. Repetition of the process using a different set of randomly selected values for each optical control parameter or action leads to strengthening of the sequence with minimal to desirable (or positive) results. Ultimately, the neural network model “learns” what adjustments to make to a set or sequence of optical control parameters or actions to achieve the target result, i.e., accurate prediction or classification.

In embodiments, the methods and systems described in this application may use convolutional neural networks (referred to in some cases as CNNs, ConvNets, shift invariant neural networks, or spatially invariant neural networks), the units of which are similar to the visual cortex of the human brain. Connected by patterns.

The initial layer of the CNN (e.g., a convolutional layer) may extract low-level features, such as edges and/or gradients, from the input object concept 720. Subsequent layers can extract or detect progressively more complex features and patterns, such as the presence of curvature and texture in the image data, etc. The output of each layer can serve as the input to a subsequent layer in the CNN to learn a hierarchical feature representation from the data in the input object concept 720. This allows convolutional neural networks to efficiently learn increasingly complex and abstract visual concepts.

In embodiments, a capsule network can be used to use fewer labeled training examples to achieve similar classification performance of a CNN.

In embodiments, a transformer-based encoder-decoder architecture using an attention mechanism may be used in conjunction with or instead of a convolutional neural network.

127 displays a schematic diagram illustrating an example implementation of a dynamic vision system showing details of various components along with integration of the dynamic vision system with one or more third-party systems in accordance with some embodiments of the present disclosure. The dynamic vision system 11900 includes a vision sensor 11904, a feedback source 11906 that provides resulting data from a machine learning system, and a system that generates data in response to changes in environmental factors including temperature, pressure, humidity, vibration, etc. to capture data from a variety of data sources 11902, including an environmental control 11908, a lighting control 11910 that generates data in response to changes in source lighting including color, color temperature, timing (PWM), amplitude, etc. It may include a configured liquid lens optical assembly 11304, and a data library 11912.

The data storage and management system 11914 allows any of the services, applications, programs, etc. to access a common data source (which may include a single logical data source distributed across disparate physical and/or virtual storage locations). Records of status and event data for various components and subsystems of the dynamic vision system 11300 may be maintained. Data storage and management system 11914 may include a memory subsystem for storage of instructions and data and a file storage subsystem that provides persistent storage for program and data files. Additionally, storage and management system 11914 may include capabilities such as data allocation, data caching, data pruning and data management, and access to and control of intelligence and data resources.

The processing system 11306 may include a machine learning system 11324 and an artificial intelligence system 11326, a digital twin system 11320, and a control system, as described in detail in Figures 123, 124, 125, and 126, or elsewhere in this application. 11314 may process data captured by liquid lens optical assembly 11304 and stored in data storage and management system 11914 to optimize and adjust optical parameters in real time.

In embodiments, a set of applications 11916 may enable dynamic vision system 11300 to present semantic information to a user and enable the user to perform specific vision tasks. Some examples of applications provided on the dynamic vision system 11918 include particle filters 11918, 3D model generation 11920, position or motion prediction 11922, visual SLAM 11924, defect detection 11926, and adversarial neural networks. Includes detection (11928).

In embodiments, dynamic vision system 11300 may include an interface, network connection, port, application programming interface (API), broker, service, connector, wrapper, container, wired or wireless communication link, human-accessible interface, software interface, It may be integrated with one or more third party systems 11930 through connectivity facilities including micro-services, SaaS interfaces, PaaS interfaces, IaaS interfaces, cloud capabilities, etc. Connectivity facilities may facilitate the transfer of data between the dynamic vision system 11300 and one or more third party systems 11930.

In embodiments, dynamic vision system 11300 may be integrated with or with a set of Value Chain Network (VCN) entities for quality control inspection and classification of objects on a production assembly line or logistics chain, where liquid lenses may be used to It is configured to quickly adjust focus to accommodate, recognize and classify objects located at a working distance or at different heights.

In embodiments, dynamic vision system 11300 may be integrated into or with a set of autonomous vehicle systems to scan the vehicle environment and monitor the vehicle's distance from other objects on the road.

In an embodiment, dynamic vision system 11300 includes an interactive head configured to display virtual content, having electrically adjustable liquid lenses to provide corrections to the displayed content by adjusting the electrically adjustable liquid lenses. It may be integrated within or with a mounting device.

In embodiments, dynamic vision system 11300 may be integrated within or with a UAV navigation system to help control the position or course of an unmanned automotive vehicle (UAV) in three dimensions.

Some non-limiting examples of third party systems 11930 that may be included in the dynamic vision system 11300 to integrate vision capabilities include IoT systems 11932, value chain network (VCN) systems 11934, and manufacturing execution systems. (MES) (11936), robot/cobot system (11938), automotive system (11940), 3D printing system, ophthalmic system, surgical system, microscopy system, exoskeleton system, prosthesis system, biometric system, quality management system (QMS) ), compliance system, certification system, etc.

In embodiments, the integration of dynamic vision system 11300 with further third-party systems 11930 takes into account the specific needs and requirements of third-party systems 11930 and, based on those requirements, dynamic vision system 11300 You can customize specific components and applications. For example, when integrating with a 3D printing system, defect detection may be provided, while integration with a robotic cleaning system may benefit from the inclusion of a virtual SLAM 11924.

128-142 relate to various embodiments of a fleet management platform configured to configure fleets of robotic operational units to perform a wide range of tasks. In some embodiments, a fleet management platform may be used as a value chain entity utilized by one or more organizations. A fleet management platform can be a standalone service or included as part of a larger multi-service offering. In an embodiment, a fleet management platform receives a work request (e.g., from a client device) and identifies a set of tasks to be performed upon completion of the requested work. In response to determining the set of tasks, the fleet management platform may determine a robot fleet configuration that includes the set of robotic operational units and assign the robotic operational units to the sets of tasks. As used in this application, a robot operating unit may refer to an individual robot, a team of robots, or a fleet of robots that operate to complete a task or set of tasks. An individual robot may refer to a special purpose robot, multi-purpose robot, exoskeleton robot, robotic process automation software, or other software-based bot, as discussed further below. As will be discussed, in some embodiments, a fleet management platform may define configurations of one or more multipurpose robots to perform respective tasks or sub-tasks and/or operate in specific types of environments as part of a fleet configuration. . As will be discussed, a multipurpose robot can be comprised of various modules that enable the multipurpose robot to perform specific tasks. For example, multipurpose robots may have specialized chips that enable the robot to perform intelligent tasks, specialized sensors for the task or environment, liquid lenses to enable specific machine vision functionality, and task-specific specialized appendages (e.g. (e.g., clamps, grippers, drilling, lifts, etc.), and/or other modules that configure the multipurpose robot to perform a specific task or set of tasks.

In some embodiments, a fleet management platform may define a set of workflows, wherein the workflows define the order in which specific tasks or sub-tasks are performed and the robot operating organization units assigned to each task or sub-task ( s) can be defined. In some embodiments, the fleet management platform may perform workflow simulations to iteratively redefine fleet configurations and/or workflows to substantially optimize the operation of the robotic fleet. For example, fleet configurations and/or workflows may be iteratively adjusted to reduce costs, improve logistics efficiency, reduce overall operation time, etc. Once the fleet configuration and workflow is complete, the fleet management platform can deploy the fleet. In some embodiments, a fleet management platform may facilitate the logistics associated with delivering robotic operating units and/or robotic components, and/or support resources to the work site(s). Additionally, in some embodiments, a fleet management platform may utilize additive manufacturing capabilities, such as 3D printers or other capabilities described in this application or in documents incorporated by reference herein, to promote resource provisioning/logistics, Items that can be 3D printed in an efficient manner can be printed instead of shipped. In embodiments, a fleet management platform may monitor a fleet of robots while performing tasks, including the status of robot operating units, performance of tasks, etc. In some of these embodiments, the fleet management platform may automate maintenance of robots and/or resources to ensure efficient use of available inventory and/or reduce downtime at work locations.

In some embodiments, the fleet management platform may include robotic operating units or edge devices, environmental sensor systems, platform resources (e.g., logistics platforms, enterprise resource management platforms, customer relationship management platforms, etc.), and/or other suitable data sources. A fleet digital twin may be supported to indicate the status and/or operational performance of a robotic operating unit based on data received from other suitable data sources, such as The digital twin served by the fleet management platform can be adapted for a variety of purposes. For example, in some embodiments, a digital twin may be configured to provide real-time status of tasks being performed by a fleet of robots. In this way, users can drill down in different areas of the job site to see progress with tasks. In some example embodiments, a digital twin may be configured to provide the status of a robot fleet, including individual robots within the fleet. In this example, the user can drill down to an individual robot within a team or fleet of robots to view the robot's status. For example, users can view the robot's battery life, availability of other energy sources, the robot's location, mobility options for the robot, the robot's productivity, the robot's task completion status, the robot's maintenance alerts, etc. In some example embodiments, a fleet management platform may serve an environmental digital twin that represents the robot fleet's environment with real-time information such as locations of objects and other robots, sensor readings of the environment, etc. In this embodiment, a user may utilize the environmental digital twin to provide remote control commands to a robot, a team of robots, or a fleet of robots. For example, a robot or team of robots may encounter an unidentified object when performing a task and may be unable to make decisions related to performing the task. In some embodiments, a fleet management platform may obtain relevant data (e.g., LIDAR data, video feeds, environmental maps, etc.) that may be displayed in an environmental digital twin. Users can view the current scenario in the environmental digital twin and command the robot fleet on how to proceed given the scenario presented in the environmental digital twin. The foregoing are non-limiting examples of digital twins that may be used in connection with a fleet management platform, and other examples are discussed below.

128 illustrates an example environment of fleet management platform 12000 (also referred to as “platform 12000”) in accordance with some embodiments of the present disclosure. In some embodiments, fleet management platform 12000 may be used as a value chain entity utilized by one or more organizations. Fleet management platform 12000 may be a standalone service or may be included as part of a larger multi-service offering. In an embodiment, robotic fleet management platform 12000 includes a fleet operating system 12002, a data processing system 12030, and an intelligence layer 12004 (e.g., platform level intelligence layer 12004). In embodiments, fleet operating system 12002 organizes and manages tasks performed by robotic operating organizational units and/or robotic operating organizational units 12040. As will be discussed, robot operating unit 12040 may refer to an individual robot, an individual robot task assembly 12050, a robot fleet 12060, and/or a robot fleet support unit 12080.

In an embodiment, fleet operating system 12002 includes communications management system 12010, remote control system 12012, resource provisioning system 12014, logistics system 12016, work configuration system 12018, and fleet configuration system. configuration system 12020, task execution system 12022, human interface system 12024, and maintenance management system 12026. In embodiments, communication management system 12010 is configured to include elements external to fleet management platform 12000 to facilitate fleet management platform communications. In an embodiment, remote control system 12012 is configured to remotely manage and enable control of robotic operational units and fleet resources. In an embodiment, resource provisioning system 12014 is configured to handle allocation and access to fleet resources (e.g., robotic operating units). In embodiments, logistics system 12016 coordinates the use and transportation of fleet resources and supplies to the shop floor and/or robotic operating organization units. In embodiments, maintenance management system 12026 facilitates coordinated and timely maintenance of fleet resources. In an embodiment, task organization system 12018 generates a task execution plan based on the task request. In embodiments, fleet configuration system 12020 configures robotic operational orchestration units (e.g., individual robots and/or robot fleets) to complete a task execution plan. In embodiments, task execution system 12022 executes, monitors, and/or reports on tasks performed by robotic operational units (e.g., according to task execution plans) to execute task plans and organize tasks and fleets. Ensures efficient use of fleet resources while addressing relevant reporting requirements. In embodiments, a human interface system provides an interface through which a human user can interface with a robotic operating unit.

As noted, robot operating organization unit 12040 may refer to an individual robot, an individual robot task assembly 12050, a robot fleet 12060, and/or a robot fleet support unit 12080. In embodiments, individual robots may include, but are not limited to, multipurpose robots 12042, special purpose robots 12044, exoskeleton robots 12046, etc. 129 illustrates a non-limiting example set of components of multipurpose robot 12100 (MPR) and special purpose robot 12180.

In an embodiment, SPR 12180 and MPR 12100 may include a baseline system 12102, a robot control system 12150, and a robot security system 12170. In an embodiment, robot control system 12140 includes a data processing system 12130 and an intelligence layer 12140. As will be discussed, the data processing system may include data processing resources that may be centralized and/or distributed among teams or fleets of robots. Additionally or alternatively, the data processing resources may include general-purpose chipsets, specialized chipsets, and/or configurable chipsets. As will be discussed, intelligence layer 12140 performs intelligence-related tasks on behalf of a robot or collection of robots (e.g., a task assembly or fleet). For example, robot-level intelligence layer 12140 may perform tasks such as artificial intelligence, machine learning, natural language processing, machine vision, analytics, etc., and may perform complex data structures (e.g., digital twins) and (e.g., from IoT, edge and other network-enabled devices, from on-premises and cloud-deployed databases and other resources, and/or from APIs, event streams, logs, or other data sources, among others) Heterogeneous data sources can be utilized. The robot-level and fleet-level intelligence layers are discussed in more detail below. In embodiments, robotic security system 12170 performs security-related functions on behalf of a robot or collection of robots (e.g., a task assembly or fleet). These security-related functions may include self-adaptive and non-adaptive security functions as well as passive security functions.

In embodiments, the baseline system 12102 of MPR 12100 or SPR 12190 includes an energy storage and power distribution system 12104, an enclosure 12106, an electromechanical and/or electrofluidic system 12108, and a transportation system. 12110, vision and sensing system 12112, and/or structural system 12114. As discussed further below, the configuration of the baseline system of SPR 12190 depends on the type of task that SPR 12190 is configured to perform. For example, the baseline system for an autonomous drone is significantly different from the baseline system for an autonomous vehicle or factory floor robot. Similarly, the baseline system of MPR 12100 depends on the type of environment in which MPR 12100 is intended to operate. For example, an MPR 12100 configured to operate in deep-sea conditions may have a different baseline system than an MPR 12100 configured to operate in arctic conditions or an aerial robot.

MPR 12100 differs from SPR 12190 in that MPR 12100 can be configured to perform a wider range of different tasks. In embodiments, MPR 12100 may further include a module system 12120 that allows MPR 12100 to be comprised of various hardware and/or software components. In this way, MPR 12100 may be equipped with different attachments, sensor sets, chipsets, synchronization adapters, etc., depending on the scope of tasks that MPR 12100 is configured to perform. In an embodiment, module system 12120 may include a control module interface 12130 and a physical module interface 12122. Control module interface 12130 and physical module interface 12122 may refer to mechanical, electrical, and/or digital interfaces that receive auxiliary components to configure MPR 12100 to perform specific tasks. In embodiments, control module interface 12130 receives (or otherwise “couples”) auxiliary components that modify one or more features associated with control of MPR 12100. These may include chipsets (e.g., AI chipsets, machine learning chipsets, machine vision chipsets, communication chipsets, etc.), sensor modules, communication modules, AI modules, security modules, computing modules, etc. In embodiments, physical module interface 12122 receives (or otherwise couples to) physical actions that may be taken by MPR 12100 and/or auxiliary physical modules that modify the physical actions of MPR 12100. Examples of physical modules may include, but are not limited to, end effectors, synchronous adapters, 3D printers, power sources, etc. As will be discussed, MPR 12100 may be reconfigured to perform one or more tasks upon completion of a task. In this embodiment, the fleet management platform 12000 may define a job execution plan and a fleet of support robots, and cause the MPRs 12100 within the fleet of support robots to be reconfigured to perform one or more specified tasks in the job execution plan. 12100), you can provision one or more modules.

Referring back to Figure 128, individual robot task assembly 12050 may refer to a collection of one or more individual robots assigned to perform a specific task or set of related tasks. The robots within the robotic task assembly may include any combination of MPR 12042, SPR 12044, exoskeleton robot 12046, etc. In some embodiments, an individual robot task assembly 12050 may include a local manager that controls or otherwise provides instructions to the robots within the task assembly 12050. In this embodiment, the local manager may be a designated supervisor robot or human operator. In embodiments, a supervisor robot may refer to a robot assigned to organize, direct, monitor, reallocate, and/or reorganize (or request reconfiguration) robots in task assembly 12050. In embodiments, the robotic supervisor may act as an edge device on behalf of the task assembly 12050 such that the robotic supervisor communicates with the fleet management platform 12000 or other suitable device or system and/or manages the task assembly 12050. may be assigned specific processing and/or communication capabilities that enable it to perform data processing operations on its behalf. In embodiments, a robot fleet is a collection of individual robots and/or task assemblies that collectively perform a set of projects upon completion of work. In embodiments, a fleet of robots may include individual SPRs, MPRs, exoskeletons, etc. Additionally, fleets may be arranged as fleets of task groups, regional fleets, and/or fleets of swarms. In embodiments, a robot fleet may be supported by robot fleet support. In embodiments, examples of robotic fleet support include buildings, edge and IoT devices, local data storage (and corresponding data interfaces), maintenance support, charging stations and devices, replacement parts, batteries, accessories, shipping containers, docking stations, May include spare parts, and/or technicians.

130 illustrates the data processing system 12030 and intelligence layer 12004 of the fleet management platform 12000. In an embodiment, data processing system 12030 includes data handling services 12032 and data processing services 12034. Data handling service 12032 is configured to store, retrieve, and otherwise manage data in fleet management platform 12000. In embodiments, data handling service 12032 accesses a set of data stores 12036 and/or libraries 12038, whereby data handling service 12032 supports fleet management platform 12000 and/or robotics operations orchestration. Write and read data from data store 12036 and/or library 12038 on behalf of other components of unit 12040. In embodiments, data processing services 12034 perform data processing operations on behalf of other components of fleet management platform 12000 and/or robotics operating organization unit 12040. For example, the data processing service 12034 may perform database operations (eg, table join, search, etc.), data fusion operations, etc. In embodiments, a data processing system may include distributed resources, centralized resources, and/or “on-chip” resources.

In embodiments, platform-level intelligence layer 12004 performs intelligence services on behalf of other components of fleet management platform 12000 and/or robot operations orchestration unit 12040. As will be discussed, in some cases platform-level intelligence layer 12004 may be configured as part of a broader intelligence system (FIG. 131), whereby decision-making and other intelligence-based functions are performed at the lowest possible level. In an embodiment, platform-level intelligence layer 12004 includes an intelligence layer controller 12007 and a set of artificial intelligence services 12005. In embodiments, artificial intelligence service 12007 may be a digital twin that manages and/or serves a set of digital twins (e.g., a robot digital twin, a robot team digital twin, a robot fleet digital twin, a logistics digital twin, an environment digital twin, etc.). Can include twin systems. In embodiments, artificial intelligence services 12007 may include machine learning (ML) systems, rule-based intelligence systems, expert systems, analytical systems, and robotic process automation, as referenced throughout this disclosure or documents incorporated by reference herein. (RPA) systems, machine vision systems, natural language processing (NLP) systems, neural network systems, and/or other intelligence or data handling systems. In an embodiment, intelligent controller 12007 includes an analytics management module, a governance library, and an analytics module.

INTELLIGENCE LAYER

131 illustrates an example intelligence layer 12200 according to some embodiments of the present disclosure. In an embodiment, intelligence layer 12200 may be a respective component of a robot-as-a-service ecosystem (e.g., platform-level intelligence layer 12004, robot-level intelligence layer 12140, or fleet-level intelligence layer (not shown)). 104 is adapted from Intelligent Services 8800 to provide a framework for providing intelligent services at this level. In such embodiments, the intelligence layer 12200 framework can be at least partially replicated at individual robots and/or at the fleet-level, such that individual robots attempt to generate decisions, recommendations, reports, instructions, predictions, classifications, etc. While the intelligence layer 12200 may be utilized to Prediction, classification, etc. can be done by the platform-level intelligence layer 12004. In this embodiment, requests for intelligence may be pushed to a higher level. For example, if a robot is not sure whether there is an occlusion in its path, the robot can escalate the request to the fleet level, where one or more additional robots can determine whether this is an occlusion in the requesting robot's path. You can work with robots. In another example, unexpected changes in the environment (e.g., changes in weather or other conditions) may cause the robot fleet level intelligence layer to change the task execution plan. In this example, the fleet-level intelligence layer may not have enough information or processing resources to safely change the job execution plan. In response, the fleet-level intelligence layer may escalate the decision to the platform 12000-level intelligence layer 12004 so that the platform 12000-level intelligence layer 12004 can plan task execution given changes in the environment. Recommended changes can be determined.

In an embodiment, intelligence layer 12200 receives requests from a set of intelligence layer clients 12260. Depending on where the intelligence layer 12200 is implemented within the robot fleet framework (e.g., fleet management platform-level, fleet-level, or robot-level), the intelligence layer client 12260 can interact with various components of the fleet management platform. Components (e.g., remote control system 12012, logistics system 12016, work organization system 12018, fleet organization system 12020, work execution system 12022, etc.), robot fleets (e.g., teams or one or more MPRs and/or SPRs within a fleet), or individual robots (e.g., a robot's robot control system, various modules of an MPR, etc.). In an embodiment, intelligence layer client 12260 provides an intelligence request to intelligence layer 12200, where the request is a specific intelligence task (e.g., decision, recommendation, report, instruction, classification, prediction, training action, NLP request, etc.). In response, intelligence layer 12200 executes the requested intelligence task.

Note that in some scenarios, the artificial intelligence service of AI system 12204 may be an intelligence layer client 12260. For example, a rule-based intelligence system may request intelligence tasks from an ML system or a neural network system, such as requesting the classification of objects and/or the motion of objects as they appear in a video. In this example, the rule-based intelligence system may be an intelligence layer client 12260 that uses classifications to decide whether to take specified actions. In another example, a machine vision system can request a digital twin of a specified environment from a digital twin system, so that a ML system can request specific data from the digital twin as features for training a machine learning model that is trained for the specific environment.

In embodiments, an intelligence task may require certain types of data to respond to the request. For example, a machine vision task is to classify objects that appear in one or more images (and potentially other data) in an image or set of images (position of the item, presence of a face, symbols or commands, expressions, parameters of motion, It requires determining features within a set of images (such as changes in state, etc.). In other examples, NLP tasks require audio of dialogue and/or text data (and potentially other data) to determine meaning or other elements of the dialogue and/or text. In another example, an AI-based control task (e.g., decisions about the movement of a robot) may require environmental data (e.g., maps, coordinates, images of known obstacles) to make decisions about how to control the motion of the robot. etc.) and/or motion planning may be required. In a platform-level example, analytics-based reporting tasks may require data from multiple different databases to generate reports. Accordingly, in embodiments, tasks that can be performed by a intelligence layer instance may require or benefit from specific intelligence layer input 12270. In some embodiments, intelligence layer 12200 may be configured to receive and/or request specific data from intelligence layer input 12270 to perform respective intelligence tasks. Additionally or alternatively, request intelligence layer client 12260 may provide specific data in the request. For example, intelligence layer 12200 may expose one or more APIs to intelligence layer clients 12260, whereby requesting clients 12260 provide specific data in a request via the API. Examples of intelligence layer inputs include sensor data (e.g., robot sensors, environmental sensors, etc.), video streams (e.g., robot-captured video streams, video camera streams, etc.), audio streams (e.g., robot-captured video streams, video camera streams, etc.), audio streams, audio streams captured from external microphones, etc.), databases (e.g., Platform 12000 databases, third party databases, etc.), human input, and/or other suitable data, but Not limited.

In an embodiment, intelligence layer 12200 may include an intelligence layer controller 12202 and an artificial intelligence (AI) service 12204. In an embodiment, artificial intelligence layer 12200 receives intelligence requests from intelligence layer clients 12260 and any necessary data to process requests from intelligence layer clients 12260. In response to the request and specific data, one or more associated services of artificial intelligence service 12204 perform intelligent tasks, and artificial intelligence service 12204 outputs an “intelligence response.” The intelligence response may refer to the output of the artificial intelligence service 12204. Examples of responses include decisions made by the artificial intelligence service (e.g. control commands, proposed task execution plans, proposed fleet configurations, proposed robot configurations, etc.), predictions made by the artificial intelligence service (e.g. text predicted meaning of a snippet, predicted outcome associated with a proposed action, predicted fault condition, etc.), classification made by an artificial intelligence service (e.g. classification of objects in an image, classification of vocal sounds, classification based on sensor data) classified fault conditions), and/or other suitable outputs of the artificial intelligence service.

In embodiments, artificial intelligence services 12204 include ML systems 12212, rule-based systems 12228, analytics systems 12218, RPA systems 12216, digital twin systems 12220, machine vision systems 12222, It may include an NLP system 12224, and/or a neural network system 12214. It is understood that the foregoing are non-limiting examples of artificial intelligence services, and that some of the systems may be included or utilized by other systems of artificial intelligence services. For example, NLP system 12224, machine vision system 12222, and RPA system 12228 may all utilize different neural networks in performing their respective functions.

In embodiments, intelligence services 12204 are ML systems that can be integrated into or accessed by fleet management platform 12000 or any fully configured robotics operational organization unit (e.g., MPR, SPR, team, fleet, etc.) Includes and provides access to (12222). In embodiments, ML system 12212 provides services to intelligent system clients 12260, such as training ML models, utilizing ML models, enhancing ML models, performing various clustering techniques, feature extraction, etc. Provide machine-based learning capabilities, features, functions, and algorithms for use by In an example, machine learning system 12026 may provide machine learning computing, data storage, and feedback infrastructure to a workflow simulation system of a task configuration system to facilitate optimizing workflow development. Machine learning system 12026 may also operate in collaboration with other fleet intelligence systems, such as rule-based systems, machine vision systems 12222, RPA systems 12216, etc.

In embodiments, artificial intelligence service 12204 may include and/or provide access to neural network system 12214. In embodiments, neural network system 12214 is configured to train, deploy, and/or utilize neural networks on behalf of intelligence layer client 12260. In embodiments, neural network system 12214 may be configured to train any suitable type of neural network that may be used by fleet management platform 12000, robots, robot teams, and/or robot fleets. Non-limiting examples of different types of neural networks include convolutional neural networks (CNNs), deep convolutional neural networks (DCNs), feedforward neural networks (including deep feedforward neural networks), and recurrent neural networks (RNNs) (including gated RNNs). (but not limited to), long-term/short-term memory (LTSM) neural networks, etc., as well as serially, in parallel, in acyclic (e.g., directed graph-based) flows, and/or intermediate decision nodes, recursive loops, etc. Any of the neural network types described throughout this disclosure and the documents incorporated by reference herein, including but not limited to hybrids or combinations of the foregoing, such as deployed in more complex flows that may include may include, where a neural network of a given type takes input from a data source or another neural network and provides output that is included in the input set of the other neural network until the flow is complete and the final output is provided. In embodiments, neural network system 12214 may be used in fleet intelligence systems, such as machine vision systems 12222, NLP systems 12224, rule-based systems 12228, digital twin systems 12226, and/or other artificial intelligence services. Can be utilized by other components. Exemplary applications of neural network system 12214 are described throughout this disclosure.

In embodiments, artificial intelligence service 12204 may provide access to and/or integrate with a robotic process automation (RPA) system 12216. RPA systems 12216 may facilitate computer automation of creating and validating workflows that involve remote control of robotic operational units, teams, fleet resources, etc., among other things. In embodiments, RPA system 12216 may monitor human interactions with various systems to learn patterns and processes performed by humans in the performance of each task. This may include observation of human actions involving interactions with hardware elements, software interfaces, and other elements. Observations include, among many other examples, field observations of humans performing actual tasks, as well as simulations or other instances of humans performing actions with the explicit intent of providing a training data set or input for the RPA system. It may involve observation of activity, for example in this case a human tagging or labeling a training data set with features that assist the RPA system in learning to recognize or classify features or objects. In embodiments, RPA system 12216 may learn to perform specific tasks based on learned patterns and processes, such that tasks may be performed by RPA system 12216 instead of or in support of a human decision maker. . Examples of RPA systems 12216 may include those of this disclosure and documents incorporated by reference herein, and may involve automation of any of the broad value chain network activities or entities described therein. In embodiments, artificial intelligence service 12204 may include and/or provide access to analytics system 12218. In embodiments, analysis system 12218 is configured to perform various analysis processes on data output from fleet functional components, such as fleet configuration system 12020, robotic operating orchestration unit, etc. In example embodiments, analytics information generated by analytics system 12218 may facilitate quantification of fleet system and system module performance compared to a set of goals and/or metrics. Goals and/or metrics may be pre-configured and dynamically determined from past fleet operation results, etc. Analysis system 12218 may be identified as performing various analytics-based processes on behalf of platform 12000, robot fleets, teams, and/or individual robots. Examples of analysis processes that may be performed by analysis system 12218 are discussed below and in documents incorporated by reference into this application. In some example implementations, the analysis process anticipates the robot's capabilities (e.g., to pick items and prepare them for delivery by autonomous vehicles, among other things) and (among other things) the demand for sets of related items, for example, by location and time. It may involve tracking goals and/or specific metrics, which may entail coordination of supply chain activities, which may entail demand intelligence.

In embodiments, a value chain network analysis system may process a set of supply chain robotic fleet data and a set of demand intelligence robotic process automation data to generate recommended actions to adjust supply and demand for a set of goods or other items. there is. In an embodiment, a value chain network automation system is provided, comprising a set of supply chain robot fleet data, comprising attributes of a set of states and capabilities of a set of robotic systems in the supply chain for a set of goods; Demand Intelligence Robotic Process Automation Data Set - containing the attributes of a set of states of a set of robotic process automation systems that perform automation of a set of demand anticipation tasks for a set of goods; and a coordination system - which provides a set of robotic task instructions for a supply chain robot fleet based on processing the supply chain robot fleet data set and the demand intelligence robotic process automation data set to coordinate supply and demand for a set of goods. Includes.

In embodiments, artificial intelligence service 12204 may include and/or provide access to digital twin system 12220. Digital twin system 12220 may include any of the wide range of features and capabilities described in this application and documents incorporated by reference herein. In embodiments, digital twin system 12220 may be configured to provide an execution environment for different types of digital twins, such as twins of physical environments, twins of robotic operating units, logistics twins, etc., among others. In an example embodiment, digital twin system 12220 includes fleet resources; Work modes, etc., such as robot operation organization units assigned to teams; It can be further configured to create digital twins for, for example, robotic operating units within a fleet. For example, digital twin system 12220 can create a digital twin of robot resources (e.g., interchangeable end effectors, power sources, communication capabilities, synchronization adapters, etc.). Additionally, the digital twin system 12220 may be configured with an interface, such as an API, for receiving information from external data sources, such as data received from the physical robot operating organization unit and/or the environment. For example, the digital twin system 12220 may receive real-time data from the sensor system of the robot operation unit and/or the sensor system of the physical environment in which the robot operation unit operates. In embodiments, digital twin system 12220 may receive digital twin data from other suitable data sources, such as third party services (e.g., weather services, traffic data services, logistics systems and databases, etc.). In an embodiment, the digital twin system 12220 may support value chain network entities such as supply chain infrastructure entities, transportation or logistics entities, containers, goods, etc., as well as demand networks such as customers, merchants, stores, points of sale, points of use, etc. It may contain digital twin data representing the characteristics, status, etc. of the entity. The digital twin system 12220 is integrated with, or linked to, an interface (e.g., a control tower or dashboard) for coordination of supply and demand, including coordination of supply chain activities and automation of demand management activities. or interact with it in some other way.

In embodiments, digital twin system 12220 may provide access to and manage libraries of robot operating unit digital twin systems. Systems, such as other artificial intelligence services 12240, may access the library to perform functions such as simulations of the actions of a robotic operating unit in a given environment performing specified tasks in response to specific stimuli. In embodiments, the digital twin system 12220 may include a robot twin (e.g., a digital twin of an individual robot operating unit), a task twin (e.g., a digital twin that can be optimized for specific task conditions/requirements, e.g. , a digital representation of a task as defined by a task definition system and/or a pre-configured library of robotic task building blocks), a team twin (e.g., a digital twin of individual robotic operational units and the tasks they are performing and/or or a digital embodiment of a designated team of robotics operational units, which may include pre-configured task-scope-specific team twins), project twins (e.g., teams of robotics operational units, which may address scopes of specific tasks). , a digital embodiment of a defined work execution plan, optionally including a digital twin for a set of tasks, fleet resources, and/or pre-configured project-specific project twins), fleet twins (e.g., maintenance, robotics operations) Aggregation of robot operating unit digital twins with fleet behavior and organizational models that take into account cross-tasking fleet functions, such as fleet decommissioning and replacement, robot operating unit backups, etc., operator twins (e.g., robotic process automation, etc.) digital embodiments of human operators, as can be determined through the use of logistics twins (e.g., digital modeling of the delivery and cost of robots, personnel, and support equipment - task-independent and addressing specific task requests) as needed), environmental twins (e.g., modeling mobility constraints and required capabilities, edge and networking constraints and capabilities, and power constraints and capabilities), as well as providing access to them. can facilitate the execution of.

In embodiments, artificial intelligence services 12204 may include and/or provide access to machine vision system 12222. In an embodiment, machine vision system 12222 is configured to process images (e.g., captured by a camera) to detect and classify objects within the images. In an embodiment, machine vision system 12222 receives one or more images (which may be single still shot images or frames of a video feed) and detects “blobs” in the image (e.g., using edge detection techniques or the like). "Identifies. Machine vision system 12222 may then classify the blob. In some embodiments, machine vision system 12222 utilizes one or more machine learned image classification models and/or neural networks (e.g., convolutional neural networks) to classify blobs within an image. In some embodiments, machine vision system 12222 may perform feature extraction on the image and/or each blob within the image prior to classification. In some embodiments, machine vision system 12222 may utilize classifications made from previous images to confirm or update the classification(s) from previous images. For example, if an object detected in a previous frame was classified with a lower confidence score (e.g., the object is partially occluded or out of focus), the machine vision system 12222 may If you can determine the object's classification with this higher confidence, you can confirm or update the classification. In an embodiment, machine vision system 12222 is configured to detect occlusion, such as an object that may be occluded by another object. In embodiments, machine vision system 12222 receives additional input to assist with image classification tasks, such as from radar, sonar, digital twins of the environment (which can show the locations of known objects), etc. In embodiments, machine vision system 12222 may output object classifications to an intelligent service client 12260, such as a robot's control system, robot supervisor, edge device, etc. In some embodiments, machine learning system 12222 (e.g., of a robotic operating unit) may include or interface with a liquid lens. In these embodiments, liquid lenses facilitate improved machine vision (e.g., when focusing at multiple distances is required by the robot's environment and tasks) and/or other machine vision tasks enabled by the liquid lenses. You can do it.

In embodiments, artificial intelligence service 12204 may include and/or provide access to a natural language processing (NLP) system 12224. In an embodiment, NLP system 12224 performs natural language tasks on behalf of an intelligence layer client 12260, such as a control system. Examples of natural language processing technologies include conversation recognition, speech segmentation, speaker diarization, text-to-speech, lemmatization, morphological segmentation, and parts-of-speech tagging. It may include, but is not limited to -of-speech tagging, stemming, syntactic analysis, lexical analysis, etc. In embodiments, NLP system 12224 may enable voice commands received from a human. In embodiments, NLP system 12224 may receive an audio stream (e.g., from a microphone) and perform speech-to-text conversion on the audio stream to obtain a transcription of the audio stream. NLP system 12224 may process text (e.g., a transcription of an audio stream) to determine the meaning of the text using various NLP techniques (e.g., NLP models, neural networks, etc.). In embodiments, NLP system 12224 may determine an action or command spoken in the audio stream based on the results of the NLP. In an embodiment, the NLP system 12224 may output the results of the NLP to an intelligent service client 12260, such as a robot's control system, a robot supervisor, an edge device, etc.

In embodiments, artificial intelligence services 12204 may also be integrated into or accessed by fleet management platform 12000 or any fully configured robotics operational organization unit (e.g., MPR, SPR, team, fleet, etc.). May include and/or provide access to a rules-based system 12228. In some embodiments, rule-based system 12228 may consist of programmatic logic that defines sets of rules and other conditions that trigger specific actions that may be performed in connection with robot fleets and/or tasks. In embodiments, rule-based system 12228 may consist of programmatic logic that receives input and determines whether one or more rules are satisfied based on the input. If the condition is met, the rules-based system 12228 determines the action to perform, which may be output to the requesting intelligence layer client 12260. Data received by the rule-based engine may be received from an intelligent data source 12270 and/or requested from another intelligence service 12204, such as a machine vision system 12222, a neural network system 12214, an ML system 12212, etc. It can be. For example, rule-based system 12228 may receive sensor data from the robot's LiDAR sensor and/or a classification of objects within the robot's field of view from the machine vision system 12222 and, in response, determine the robot's path. You can decide whether you should continue, change course, or stop. Rule-based system 12228 may be configured to make other appropriate rule-based decisions on behalf of each client 12260, examples of which are discussed throughout this disclosure. In some embodiments, the rules-based engine may apply governance standards and/or analysis modules, which are described in more detail below.

In an embodiment, artificial intelligence service 12204 is configured to interface with intelligence layer controller 12202 and determine the type of request issued by intelligence layer client 12260, and in response thereto, artificial intelligence service 12204 when responding to the request. The set of governance standards and/or analytics to be applied by the intelligence service 12204 may be determined. In an embodiment, intelligent layer controller 12202 may include an analytics management module 12206, a set of analytics modules 12208, and a governance library 12210.

In an embodiment, intelligence layer controller 12202 is configured to determine the type of request issued by intelligence layer client 12260 and, in response, the governance standards to be applied by artificial intelligence service 12204 when responding to the request. and/or determine a set of analyses. In an embodiment, intelligent system controller 12202 may include an analytics management module 12206, a set of analytics modules 12208, and a governance library 12210. In embodiments, analytics management module 12206 receives requests for artificial intelligence services and determines governance standards and/or analytics associated with the requests. In embodiments, analysis management module 12206 may determine governance standards to apply to a request based on the type of decision requested and/or whether a particular analysis should be performed for the requested decision. For example, a request for a control decision to move a robot to a different location may be tied to a specific set of applicable governance standards, such as safety standards, legal standards, quality standards, etc., and/or risk analysis, safety analysis, engineering One or more analyzes may be associated with control decisions, such as analysis, etc.

In some embodiments, analytics management module 12206 may determine governance standards to apply to a decision request based on one or more conditions. Non-limiting examples of these conditions include the type of decision being requested, the jurisdiction in which the robotic fleet is located, the geographic location in which the robotic fleet is deployed, the environment in which the robotic fleet and/or robotic operating unit is operating, and the current or predicted conditions of the environment. It may include environmental conditions, etc. In embodiments, governance standards may be defined as a set of standards libraries stored in governance library 12210. In embodiments, a standard library may define conditions, thresholds, rules, recommendations, or other suitable parameters under which decisions may be analyzed. Examples of standard libraries may include legal standard libraries, regulatory standard libraries, quality standard libraries, engineering standard libraries, safety standard libraries, financial standard libraries, and/or other suitable types of standard libraries. In an embodiment, governance library 12210 may include an index that indexes specific standards defined in each standard library based on different conditions. Examples of conditions may be jurisdictions or geographic areas to which a particular standard applies, environmental conditions to which a particular standard applies, types of robots to which a particular standard applies, materials or products to which a particular standard applies, etc.

In some embodiments, analytics management module 12206 may determine an appropriate set of standards that should be applied in connection with a particular decision, and cause artificial intelligence service 12204 to utilize the associated governance standards in determining the decision. can be provided to the artificial intelligence service (12204). In such embodiments, artificial intelligence service 12204 may be configured to apply standards in the decision-making process such that decisions output by artificial intelligence service 12204 are consistent with associated governance standards. For example, when operating a robot fleet in a particular jurisdiction or geographic area, there are certain restrictions on the types of robots (e.g., no drones), preservation of certain species or ecosystems (e.g., protected wetlands), etc. Legal or regulatory standards may be attached. In this example, decisions regarding fleet composition may exclude certain types of robots from the fleet composition (e.g., no drones) and ensure that none of the robots within the fleet pose a threat to the ecosystem in which the robot fleet will operate. It can be guaranteed. In another example, the robot's control system may request control decisions from the robot's intelligence layer. In this example, the presence of a human or other living entity in proximity to a robotic operating unit may be associated with a set of standards (e.g., safety standards, legal standards, etc.). In this example, intelligence layer 12200 may receive input such as video feeds, LIDAR data, etc. AI service 12204 may initially classify the object in analysis management module 12206 and may receive input from the video feed indicating that a human is within the robot's field of view. In response, analytics management module 12206 may determine that a particular safety standard applies and provide associated governance standards from a safety standards library to AI service 12204, which may in turn use intelligent system input (e.g. , current location, destination, video input, LIDAR data, etc.) and associated safety criteria, an attempt can be made to determine control decisions. If the AI service 12204 is unable to make a decision under a given safety standard, the AI service 12204 makes a default decision (which may be defined in a safety standards library), such as suspending and/or relinquishing control to the human operator. can be issued. It will be appreciated that the standard libraries within the governance library may be defined by platform 12000 providers, customers, and/or third parties. Standards may be government standards, industry standards, customer standards, or other suitable sources. In embodiments, each set of standards may include a set of conditions that associate each set of standards such that the conditions can be used to determine which standard to apply in a given situation.

In some embodiments, analysis management module 12206 may determine one or more analyzes to be performed for a particular decision and provide corresponding analysis modules 12208 to artificial intelligence service 12204 to perform these analyses, , the artificial intelligence service 12204 utilizes the corresponding analysis module 12208 to analyze the decision before outputting the decision to the requesting client. In embodiments, analysis module 12208 may include modules configured to perform specific analyzes on specific types of decisions, whereby each module is configured to be processed by a processing system hosting an instance of intelligence layer 12200. It runs. Non-limiting examples of analysis modules 12208 include risk analysis module(s), security analysis module(s), decision tree analysis module(s), ethics analysis module(s), failure mode and effects (FMEA) analysis module(s), s), risk analysis module(s), quality analysis module(s), safety analysis module(s), regulatory analysis module(s), legal analysis module(s), and/or other suitable analysis module(s). .

In some embodiments, analysis management module 12206 is configured to determine which type of analysis to perform based on the type of decision requested by intelligent system client 12260. In some of these embodiments, analysis management module 12206 may include an index or other suitable mechanism to identify a set of analysis modules 12208 based on the requested decision type. In this embodiment, analysis management module 12206 can receive a decision type and determine a set of analysis modules 12208 to run based on the decision type. Additionally or alternatively, one or more governance standards may define when certain analyzes are performed. For example, engineering standards may define which scenarios require FMEA analysis. In this example, engineering standards may be tied to a request for a specific type of decision (e.g., a fleet configuration request), and engineering standards may be tied to a request for a specific type of decision (e.g., a fleet configuration request), and the engineering standard may be tied to , when operating in a specific type of environment such as an enclosure, or when working with hazardous materials). To continue this example, the rule-based system 12228 of AI service 12204 may determine that the request corresponds to one of the defined scenarios, and then call the FMEA analysis module to perform analysis on the requested decision. can do.

When artificial intelligence service 12204 is performing an intelligence task involving analysis, artificial intelligence service 12204 may execute corresponding analysis module(s) to analyze potential decisions determined for the requested intelligence task. . If none of the associated analysis modules 12208 flag the decision as violating the respective analysis, the artificial intelligence service 12204 may output the proposed decision to the intelligence client 12260. If a proposed decision is flagged by one or more analysis modules 12208, artificial intelligence service 12204 may determine alternative decisions and execute the associated analysis module(s) until a decision is reached. .

In embodiments, analysis module 12208 may utilize one or more standards defined in one or more standard libraries stored in governance library 12210. In some embodiments, a standard library may define conditions, thresholds, rules, recommendations, or other suitable parameters under which decisions can be analyzed. Examples of standard libraries may include legal standard libraries, regulatory standard libraries, quality standard libraries, engineering standard libraries, safety standard libraries, financial standard libraries, and/or other suitable types of standard libraries. In embodiments, each standard library may include an index that indexes specific parameter sets defined in each standard library based on different conditions. Examples of conditions may be jurisdictions or geographic areas to which a particular standard applies, environmental conditions to which a particular standard applies, types of robots to which a particular standard applies, materials or products to which a particular standard applies, etc. In such embodiments, the analysis management module 12206 may determine the appropriate set of standards that should be applied for a particular decision, whereby the corresponding analysis modules are parameterized with the determined standards, such that the parameterized analysis module 12206 is configured to follow these standards. Each analysis is performed using . In such embodiments, analysis module 12208 may be configured to apply different standards to the same analysis based on one or more conditions surrounding the decision.

In one example, before the output of a proposed control decision directing the robot to move forward is provided to the robot's robot controller, the robot's intelligence service 12204 may request a proposed set of standards and/or rules corresponding to the control decision. Decisions can be analyzed. In this example, artificial intelligence service 12204 may execute a safety analysis module and/or a risk analysis module and determine alternative decisions about whether an action would violate a legal or safety standard. In another example, before a fleet configuration proposal is output to a requesting client, the intelligence service 12204 of the fleet management platform 100 determines that the proposed fleet configuration does not violate any jurisdictional legal or regulatory standards (e.g., types of robots may be prohibited from operating in certain areas or environments, certain communication protocols may be prohibited in certain areas or environments) and/or do not potentially jeopardize the quality of task performance (e.g., selected Configurations may include robots that do not perform well in certain conditions) and/or operating certain types of robots in unsuitable conditions, such as freezing temperatures, areas of high humidity, salt or fresh water, etc. ) can be analyzed to ensure that the proposed fleet configuration. In response to analyzing the proposed decision, artificial intelligence service 12204 may optionally output proposed conditions based on the results of the performed analysis. If the decision is accepted, artificial intelligence service 12204 may output the decision to requesting intelligence layer client 12260. If the proposed configuration is flagged by one or more of the analyzes, artificial intelligence service 12204 may determine alternative decisions and run the analysis on the alternative proposed decisions until a matching decision is obtained.

Note here that, in some embodiments, one or more analysis modules 12208 may themselves be defined as standards, and one or more associated standards used together may include specific analyzes. For example, applicable safety standards may require a hazard analysis or more acceptable method to be used. In this example, ISO standards for overall processes and documentation, and ASTM standards for narrowly defined procedures may be used to complete the risk analysis required by the safety governance standard.

As mentioned, the above-described framework of intelligent system 12200 can be applied at various levels of the disclosed environment. For example, in some embodiments, a platform-level intelligence system (e.g., intelligence layer 12200) may consist of the overall capabilities of intelligence system 12200, and specific configurations of intelligence system 12200 may be used to operate the robot. Provisioning may be made for each robot operating organization unit according to the tasks assigned to the organization unit. Additionally, in some embodiments, the robotic operational organizing unit may perform intelligent system tasks at a higher level (e.g., fleet level, edge device, or platform 12000) when the robotic operational organizing unit is unable to perform the task autonomously. It can be configured to be updated with . Note that in some embodiments, intelligence layer controller 12200 may direct intelligence tasks to lower level components. For example, the intelligence layer controller 12202 of the robot fleet or fleet management platform 12000 may request the intelligence layer 12200 of a particular robot if the robot has access to the intelligence data source 12270 required for the intelligence request. Intelligence requests can be directed. Additionally, in some implementations, intelligence layer 12200 may be configured to output a default action when a decision cannot be reached by intelligence layer 12200 and/or a higher or lower level intelligence layer. In some of these implementations, default decisions may be defined in rules and/or standard libraries.

SECURITY SYSTEM

132 illustrates an example of a security system 12280 according to some embodiments of the present disclosure. In embodiments, security system 12280 illustrates a framework that can be implemented at various levels of the disclosed system. In these embodiments, instances of security system 12280 may be implemented at platform 12000-level, fleet- or team-level, or individual-level. For example, at the platform 12000 level, security system 12280 provides security-related functionality in connection with any communication and/or other interaction on behalf of platform 12000 and/or with a robotic operating unit. can do. In embodiments, a security system 12280 implemented at the fleet-level or team-level may enable the security system to communicate and/or otherwise interact with robots on behalf of and/or within a team or fleet of robots. It may be configured to provide security-related functionality. In embodiments, the security system 12280 implemented at the robot-level may be configured to provide security-related Can be configured to provide functionality.

In embodiments, security system 12280 may include an autonomous adaptive security module 12282, an autonomous non-adaptive security module 12284, and/or a passive security module 12286. The autonomous adaptive security module 12282 may be configured to request an intelligence task from the intelligence layer 12200, whereby the adaptive security module 12282 utilizes the artificial intelligence service 12204 of the intelligence layer 12200. Evaluate security risks and determine actions based on the output of the intelligence layer 12200. For example, the adaptive security module 12282 of a robot fleet may identify potentially dangerous One or more conditions associated with a robot fleet can be monitored by receiving data from a set of data sources, such as monitoring a work area for conditions. In response to receiving the data, adaptive security module 12282 may request an assessment (e.g., classification) of the environment from intelligence system 12200 regarding the security of the environment. In response, intelligent system 12200 may provide one or more classifications indicating an assessment of the environment. Adaptive security module 12282 can then determine whether the evaluation requires action to be taken and, if so, what specific action to take. In some of these embodiments, adaptive security module 12282 may determine whether an evaluation is necessary and what action to take, and, if necessary, what action to take using a rule-based approach. Additionally or alternatively, adaptive security module 12282 may be configured to detect a given set of characteristics (e.g., classification, sensor readings from one or more robots, the location of the robot, detected objects and their locations in the environment, and/or any You can utilize a neural network that is trained with actions to recommend other relevant features. In this embodiment, neural network system 12214 may receive features from adaptive security module 12282 and/or a set of intelligence layer inputs 12270 and, given the set of features, output suggested actions. there is. In some of these embodiments, intelligent controller 12202 of intelligent system 12200 may accept or override decisions made by artificial intelligence service 12204. For example, analysis module 12208 may perform dynamic risk analysis 12292 and/or static risk analysis 12294. Examples of dynamic risk analysis include real-time data-driven analysis (e.g. current weather patterns, current political region, current health crisis, etc.) and/or task-specific risk analysis (e.g. contractual risk, environmental risk, safety liability). , financial liability, etc.), but is not limited to this. Examples of static risk analysis may include, but are not limited to, operational risks (e.g., product design risks, manufacturing risks, quality control risks, etc.) and/or regulatory/compliance risks.

In embodiments, autonomous adaptive security module 12282 operates in an isolated manner (e.g., without communication with an external device or system) or in a connected manner (e.g., with communication with an external device or system). can do.

In embodiments, security system 12280 may include autonomous non-adaptive security module 12284. In embodiments, autonomous non-adaptive security module 12284 is configured to make security-related decisions autonomously (e.g., without human intervention) on behalf of a client. In embodiments, non-adaptive security module 12284 performs logic-based security-related actions (e.g., risk mitigation actions) in response to detecting one or more specific sets of conditions. For example, non-adaptive security module 12284 may, in response to detecting a particular set of conditions, turn off the robot, stop the robot's movement, initiate charging, sound an alarm, or other actions. It may be configured to trigger actions, such as sending a notification to the device or system, self-destructing, etc. In embodiments, non-adaptive security module 12284 responds to more easily diagnosable hazards, such as overheating conditions, movement or being pulled out of a geofenced area, detected internal leaks, low power conditions, low fluid levels, etc.

In an embodiment, security system 12280 may include a passive security module 12286. In embodiments, passive security module 12286 is configured to enable users to make decisions regarding security-related actions. In some of these embodiments, passive security module 12286 may receive notification of assessed risk (e.g., from adaptive security module 12282, non-adaptive security module 12284, intelligent client 12260, etc.). configured to receive. In such embodiments, a human user may interface with passive security module 12286 through a human interface, which may be provided through a user device (e.g., mobile device, tablet, computing device, etc.).

Various security and risk mitigation strategies are discussed throughout this disclosure.

133 illustrates an example set of components of a fleet operating system 12002 of a fleet management platform. In embodiments, fleet operating system 12002 may utilize features and functionality of robotic fleet management platform 12000 to facilitate substantially optimized use of fleet resources by anticipating fleet resource demands and preparing those resources in advance of anticipated use. You can use your abilities. In embodiments, resource demand forecasting may include coordinating maintenance activities with job scheduling to ensure preventable outages due to lack of maintenance are avoided. Additionally or alternatively, resource demand projections may be based on alignment of detected fleet resource usage with information supporting projections of work requests, among other things. In embodiments, factors such as expected weather patterns, time of year, location, etc. may affect the likelihood of a particular work request (e.g., during hurricane season, urgent infrastructure repair work is likely to be requested). An example implementation for generating fleet demand forecasts and addressing such forecasts follows a discussion of the components of the fleet operating system 12002 and the associated robotic fleet management platform 12000. As previously discussed, example components of the fleet operations system 12002 include a communications management system 12010, a remote control system 12012, a resource provisioning system 12014, a logistics system 12016, and a work organization system 12018. ), a fleet configuration system 12020, a job execution, monitoring, and reporting system 12022 (also referred to as a “job execution system” 12022), and a human interface system 12024.

In an embodiment, communications management system 12010 includes a fleet operating system 12002 and elements thereof as described herein, a fleet intelligence layer 12004 and elements thereof as described herein, and external data sources 12036. ), communication (e.g., efficient and/or high-speed communication) between fleet management platform elements such as third-party systems (e.g., via the Internet, etc.), robotic operating orchestration units, support systems and equipment, human fleet resources, etc. ) is configured to enable. Communications management system 12010 may include or provide access to one or more types of communications networks, such as wired, wireless, etc., which may support various data protocols such as Internet Protocol (IP). The communications management system includes intelligence services (e.g., through the Fleet Intelligence System resources described herein) that manage and control portions of the fleet management platform infrastructure associated with communications to ensure, for example: or have access to: Timely delivery of data collected by deployed robotic operational units to critical computational, analytical and/or data storage resources; Prioritized delivery of robot configuration and motion commands; etc. Fleet Resource Management and Control In an embodiment, communications management system 12010 prioritizes the use of fleet security system communications of fleet communications resources over communications between fleet intelligence system components to support a high degree of security and integrity of fleet resources. You can get angry. Communications management system 12010 includes at least a fleet platform network system (e.g., fleet edge device, etc.) that connects fleet management platform 12000 with external systems, deployed robotic operating units, and other network-connectable elements (e.g., fleet edge devices, etc.) 380) can provide and manage access to networking.

In embodiments, the capabilities of the communications management system 12010 may include, among other things, robot fleet communications resources (e.g., networks, radio systems) based on task execution plans, plan definitions, task definitions, robot operation orchestration unit configurations, real-time task status, etc. , data communication devices such as routers, etc.), context specifications, and/or adaptations. Communications Management System 12010 adaptation of fleet communications resources is influenced by various real-world conditions (e.g., weather, atmospheric conditions, building structures, working environment (e.g., land-to-submersible, underground), etc.). You can receive it. In embodiments, communication management system 12010 may obtain context from task requests that may facilitate anticipating the need for and type of adaptation during task execution. As an example of task request context-based communication adaptation, a task may begin at sea level and then include actions by underground teams and high-altitude teams. Communication resources suitable for use in these different task environments that are configured by the fleet configuration system during task configuration activities can be adapted by the communications management system 12010 for each team of robots as the task progresses through the exemplary environment. can be controlled.

Task request criteria can directly invoke isolated actions. Alternatively, the circumstances of the work request may favor isolated operation (e.g., operation within a foreign jurisdiction, etc.). Communication resources for the requested task can be adapted accordingly. By way of example, communications between teams of fleet resources assigned to collocate when performing a task may be configured by the fleet configuration system with additional encryption or radio frequencies to defend against conventional detection, thereby allowing the communications management system to This can facilitate activation when required by request (for example, when a team enters a foreign jurisdiction as mentioned above). Additionally, communications outside of the team may be limited to specific sites by the communications system, such as only when the entire team is located outside of a high-risk area or other designation (e.g., within a building, etc.). In this example, the delivery robot may be configured to travel from a collocated operation site to a secure external communication site to exchange information with a remote fleet management facility, etc., and upon return to the collocated site, may be configured to process authorized communications for that location and You can only use the system. This non-limiting example illustrates a representative degree of diversity in communication capabilities and conditions that will be handled by a fleet communication management system. Isolated operation does not include inter-robot operating unit communication, such as wireless communication, in addition to or instead of meeting task request requirements and/or environmental constraints (e.g., working in a remote mountain or other isolated environment). It may not be possible. In this additional embodiment of fleet resource configuration, communications management system 12006 may be configured to use communications resources (e.g., robotic operating organization unit radio interfaces, communications infrastructure proximate to an isolated robotic operating organization unit, etc.) to effectuate such fleet configuration. ) can be detected and controlled. Another consideration for isolated operation is to only allow the use of low-energy short-range communications conditionally based on the deployment situation, such as the expected location of the team robots, for example when multiple robot operating units are expected to be nearby. It may include adaptable isolation communication protocols such as Communication system 12006 is responsible for configuring robot operating units, selecting robot units that meet task request communication requirements, and configuring and specifying the deployment of fleet communication resources (e.g., teams and robots). The fleet configuration system can be assisted with fleet configuration, such as juxtaposing relay devices between operating units, and other fleet and robot configuration considerations. In this example of fleet configuration assistance, a work request may indicate a preference for using a particular robot operating unit. The fleet configuration system may query the communications control system regarding adaptation capabilities (e.g., the fleet communications management system and/or specific fleet communications resources) to support the preferred robot operating organization unit.

In an example of a communications management adaptability capability to support various robotic operating organization unit communications configurations, communications management system 12010 may be configured to configure a robotic operating organization unit that is performing field operations in the same radio signal range as a first team of robotic operating organization units. capable of supporting a first team of robot operation organization units that perform field operations using a different radio frequency for wireless communication than the second team of; Thereby, the possibility of cross-radio interference can be mitigated. Additionally, communications management system 12010 may provide reliable communications through the use of redundancy, such as automatic channel selection (e.g., local networking, cellular networking, mesh networking, long-range satellite networking, etc.), through a dual radio system. You can. Fleet communication resources may include robot operating units that act as network elements, for example, when the robot operating units are comprised of one or more mesh networks, etc. The robotic operating unit may facilitate communication in other ways, including visually, such as through the use of a light source (e.g., Morse code or binary transmission), physical gestures, infrared signals, etc. Auditory communication between robots (e.g., non-human language encoded audio signaling), ultrasound, and other auditory-based technologies may be rendered as forms of communication between robots. Much like juxtaposed robots from different teams may use different radio frequency signals, juxtaposed robots may use different auditory signaling to aid in communication clarity between team members.

In embodiments, communication management system 12010 may be configured as a plurality of independent communication systems configured to meet at least corresponding portions of a fleet's communication needs. In an example, communications management system 12010 is a first communications system for communicating between elements (or any other fleet system, system, module, team, fleet segment, etc.) within fleet operating system 12002 and fleet intelligence. may consist of a second communication system for communicating between hierarchical elements (or any other part of the fleet platform that may be separate from the first communication system), so that disruption of any individual communication system is isolated from other platform communication systems. This may reduce the impact of communication problems across platform 12000. Additionally, in this example, the fleet operating system 12002 and its components (e.g., work organization system 12018, etc.) may provide access to fleet intelligence layer elements, such as machine learning systems, to the fleet intelligence layer serving 2 Even if the communication system may be compromised due to a problem, it will still be able to communicate via the primary communication system and perform virtually all relevant fleet operational functions (including communication with remotely deployed fleet robot operating organization units, etc.). there is. Additionally, communications management system 12010 may include security features that effect isolation and shunning of platform systems, systems, system elements, communications systems, and other platform resources that appear to have been compromised by malware, etc. Other independent communication systems include robot-to-robot communication systems, human-to-robot communication systems, emergency response communication systems, etc. Another independent communication system may be based on aspects such as confidentiality of information (e.g., negotiation between fleet management provider and task requester), fleet operation oversight, etc. In embodiments, communication management system 12010 may be configured to provide role-based (etc.) access to different communication systems. As an example, a task execution system executing a first requested task may not be provided access to a particular resource based on geofence conditions (eg, when the resource is outside a designated area). In another example, a fleet operations executive may be granted concurrent access to robotic operations organizational units assigned to different tasks to perform fleet oversight functions.

In addition to and/or instead of separate communications systems, fleet communications management system 12010 may provide redundancy (multi-frequency radio, etc.) to address exceptional conditions that may cause network damage, emergency use, etc. It may be requested to override the operational communication channel for this purpose.

In an embodiment, the fleet communication management system 12010 may configure a fleet resource-specific (e.g., For example, individual robot operating units) can provide secure communications. Fleet communications management system 12010 may additionally provide broadcast capabilities to support notifications, updates, alerts, and other services. Broadcast capabilities can be fleet-wide (e.g., notification to all fleet resources to observe daylight saving time), or team-specific (e.g., all teams about a team member's role change). updates to members), task-specific (e.g., an alert that a task is on hold, a fleet resource assigned to the task - which may include multiple robot teams), or a specific type of fleet resource (e.g. For example, a fleet robot operating unit, a multipurpose robotic operating unit, one or more types of special purpose robotic operating units, such as a robotic operating unit configured with a supervisory role), a fleet support unit, a location-specific unit (e.g. , all units within a flash flood zone), may be fleet resource type-specific, etc.

In embodiments, fleet communications management system 12010 may utilize a fleet security system 12006, a fleet network system 380, and various resources, including artificial intelligence (AI) chipsets, data encoders, communications spectrum frequencies, etc. Task-specific communication elements may be used or managed in conjunction with other fleet management platform features or services, including without limitation. Fleet communication management system 12010 may operate in conjunction with fleet security system 12006, such as by providing secure high-up-time access to fleets and associated communication resources. As an example, fleet security system 12006 may utilize portions of configured communication channels (e.g., wired computer-to-computer links, wireless networks, etc.) that may be reserved by a communication management system for secure use. Some of these may include physically dedicated elements (e.g., wired connections, wireless access points operating over a dedicated set of frequencies, etc.). In embodiments, providing dedicated wireless access may include prioritizing secure system access to an existing wireless network, such as by routing secure system data packets, streams, etc., ahead of non-secure system packets. You can. As another example, a communications management system may allocate communications devices with greater battery energy (higher charge) and/or fixed power for secure system use, while lower power, lower energy, for non-secure system use. , and/or rechargeable devices may be assigned. Security system communications resource management and control may be fleet-wide, task-specific, team-specific, deployment-local based, geo-location-based, etc. By way of example, a fleet configuration system may specify configuration of fleet communication resources to meet the security aspects of a requested task. This configuration can be applied to fleet resources and maintained by the communications management system for the duration of the resource's participation in the requested task.

Additional cooperative operation of the fleet security system 12006 with the fleet communication management system 12010 may prevent access by the fleet resources to external resources (e.g., websites, etc.) as well as access by external resources to the fleet resources. This may include management. The fleet security system 12006 may deploy security agents, etc. to fleet resources based on the allocation/configuration of those resources. By way of example, the firewall type security functionality of fleet security system 12006 may be deployed, among other things, at access points managed by the fleet communication management system to connect separate task specific communication systems. Fleet communication system 12010 may also coordinate mobile robot operations configured by a fleet configuration system to have access to multiple isolated communication systems (e.g., a hub-type arrangement to facilitate access between isolated communication systems). Can support the management of one or more fleet resources, such as units. Fleet security system 12006 can enforce access rights between communication systems by deploying and operating centrally managed threat detection and management system agents in these hubs.

In an embodiment, the fleet communications management system 12010 is configured to enrich fleet security capabilities and establish dynamic communications management capabilities that work in conjunction with fleet security capabilities to further reduce the likelihood of successful intrusion into a fleet communications system, including: It can take advantage of the intelligence capabilities of fleet resources, such as resources with artificial intelligence capabilities (optionally provided by AI-specific chips and chipsets, etc.). For example, relying on AI-based functionality deployed across at least portions of fleet resources (e.g., individual robotic operating units, etc.) to infiltrate (e.g., based on context and historical information representative of such environment, etc.) It may be possible to detect local environments where there is an increased risk of other threats, so the communications management system, optionally in conjunction with the fleet security system 12006, can adapt fleet communications resources to reduce such risks.

Fleet communication management system 12010 may facilitate control of and/or use of other fleet network systems 380. As an example of management of the fleet network system 380, the fleet communication management system 12010 may, for example, determine which resources utilize the network, how resources using the network at the same time can be coordinated, and limit network loading for those resources. By determining and/or controlling the like, the fleet network system 380 can be treated to be managed as a resource for use as a fleet resource for communication.

In embodiments, the fleet operating system 12002 is configured to assist the job execution system 12022 and provide a framework for remotely controlling robotic operating units and other external resources to complete jobs and/or tasks. Includes control system 12012. In embodiments, remote control system 12012 may define and use control signals for remote operation of robotic operational units (e.g., multipurpose, special purpose, exoskeleton, human, etc.), fleet support units, external resources, etc. It can be managed. Robot remote control, as enabled by remote control system 12012, may include local robot operating unit to robot operating unit control signaling, such as when a team supervisor robot is instructing one or more robot team members to perform a task. It may include the definition and management of . Other examples of remote control signal management may include human-to-exoskeleton signaling, robot-to-robot fleet support signaling, intra-team robot operation unit signaling, etc.

In embodiments, fleet operating system 12002 is configured to assist task execution system 12022 and provide a framework for remotely controlling robotic operating units and other external resources to complete tasks and/or tasks. Includes control system 12012. The remote control system may manage the definition and use of control signals for remote operation of robotic operating units (e.g., multipurpose, special purpose, exoskeleton, human, etc.), fleet support units, external resources, etc. Robot remote control, as enabled by remote control system 12012, may be performed on a local robot operating organization unit to a robot operating organization unit, e.g., when a team supervisor robot is directing one or more robot team members to perform a task. May include definition and management of control signaling. Other examples of remote control signal management may include human-to-exoskeleton signaling, robot-to-robot fleet support signaling, intra-team robot operation unit signaling, etc. In embodiments, the remote control system may be configured to include, for example, a fleet configuration and/or fleet management platform 12000, including a fleet communications management system 12010, a fleet security system 12006, and/or a fleet network system 380. ) resources to access information, in some cases make decisions, and execute commands. A framework for remotely controlling a robotic operating unit may include a set of action-based standard rules, adapted rules modified by situational awareness, emergency rules, exceptions, human decisions, ethical rules, fleet intelligence systems, etc. However, specialized, fallover, or other communications required to handle a range of remote control requirements may be part of a communications management system 12010 that can facilitate the delivery of remote control communications/signaling, while also allowing may be determined from the use of remote control system 12012.

Remote control system 12012 may include a local supervisor remote control initiator, a human (local or remote) remote control initiator, an automated fleet-based remote control initiator (e.g., a fleet artificial intelligence system, etc.), (e.g. , law enforcement, etc.) may recognize multiple initiators of remote control signals, including third party remote control initiators. Remote control signaling may include managing remote control signals to resources external to the fleet, such as fire and emergency response resources, infrastructure resources, third party robotics service providers, etc.

Fleet resources that can participate in remote control operations can vary in both implementation and protocol, such as older generation robotic operational units, human fleet resources, quantum computing elements, etc. Accordingly, remote control system 12012 (cooperating with communication system 12006) may use multiple remote operation protocols (multi-protocol) to ensure that any two devices exchanging control signals can do so reliably. ) can be comprised of knowledge of ability. In embodiments, multi-protocol capabilities may include handling and/or providing services as protocol-to-protocol conversion, remote control signal integration and interpretation, protocol normalization, etc. In embodiments, communications management system 12010 may be configured with such protocol handling capabilities (e.g., by being deployed in conjunction with the protocol handling capabilities of remote control system 12012), either directly as noted above and by API, etc. ) You can use these protocol handling capabilities. In embodiments, remote control system 12012 (or equivalent functionality integrated with communications management system 12010) may utilize digital twins and/or artificial intelligence, e.g., to facilitate protocol conversion and/or adaptation. It can rely on parts of the fleet intelligence layer such as services. Accordingly, remote control system 12012 may optionally provide real-time, on-demand protocol conversion supported by a fleet intelligence system. Remote control system 12012 may support fleet external remote control via ports configured for integration with external and/or third party remote control architectures. Remote controls may be communicated via dedicated infrastructure and/or communication features (eg, short-range broadcast capabilities).

The remote control system 12012 may include ethical capabilities that can provide guidance and/or regulation of remote control based on ethical factors, such as ensuring that the robot does not cause harm to humans, animals, the environment, etc. there is. Ethical factors may be influenced by government and/or industry regulations, models of human behavior that facilitate determining fairness, etc. Ethics can be enforced through statistical measures, such as based on voting by members of a team of robot operating units. As an example of statistical-based ethics enforcement, an action to override a work execution plan, an attempt at a remote handover of a robotic operating unit, or any other exception may be evaluated by a portion of the team members, where each portion of the portion Members can contribute perspectives on remote operations. Each view can vote to allow/acquire a remote control action. The robot operating unit vote may be split among possible outcomes (e.g., 90% yes, 10% against), etc. to establish some form of weighting of views on the possible outcomes. The remote control system 12012 may be configured to be influenced by ethics-based decision making, such as robot operating organization unit voting as described herein. Ethics-based control, etc. can be combined with other remote control system 12012 control capabilities so that factors other than ethics, such as cost, etc., can be considered in remote control. In embodiments, ethical capabilities may be utilized through intelligence layer 12200. In such embodiments, remote control commands may be analyzed using one or more analysis modules 12208 and/or with respect to one or more sets of governance standards.

Remote control, such as control of a robotic operating unit, may be initiated at least in part by a human operator. In embodiments, fleet operating system 12002 may encounter unexpected and/or unknown conditions during task execution (e.g., as may exemplarily be reported by task execution system 12022); Remote control of the robotic operating unit(s) can be deferred to a human operator. Optionally, one or more fleet intelligence system 12022 components, such as an artificial intelligence system, may be referenced for at least a candidate remote control signal. In embodiments, the task execution plan may indicate that, in predetermined motion tasks, robot motion should be guided by a human operator. When such a task is expected to occur in a task workflow (e.g., by a task execution monitoring instance, such as a supervisor robot, etc.), the robot executing the workflow invokes human operator control with the appropriate human operator, the robot operating orchestration unit. , a remote control system 12012 may be invoked to supervise remote control connections between teams, team supervisors, etc.

In embodiments, remote control system 12012 may have access to a set of remote control signal sequences to remotely perform specific tasks. System 12012 may suggest one or more remote control signal sequences to human operators and/or automated control systems based on the context of the workflow being performed. In embodiments, the remote control system may receive input from a human operator (e.g., commands such as “stop,” “evacuate,” etc.), optionally with the assistance of other fleet resources (e.g., artificial intelligence systems, etc.). A set of remote control signals may be generated for processing and remotely controlling fleet resources, such as robotic operating units. Remote control signal sequences can be pre-configured to handle a variety of real-time situations such as security breaches, equipment failures, etc. In addition to facilitating and/or managing remote control of robotic operational units, remote control signal sequences may be used for reconfiguration of deployed and/or allocated fleet resources for tasks, workflows, operations, etc. In an example of the use of remote control signals for reconfiguration, a set of robotic operational units performing a task may be remotely controlled to assume a new role due to the failure of one of the robots in the set. A human operator (or an automated system monitor-type application) may, for example, communicate remote control signals to one or more of the executable members of the team to communicate with a robot operating unit configuration server to receive reconfiguration instructions and reconfiguration data. Provides remote control signals communicated to executable members to adjust task roles and actions accordingly.

Although generally described in this application as remote control signals, remote control system 12012 converts remote control signals into remote control instructions (e.g., combinations, abstractions, etc. of remote control signals) at the fleet level, team level, robot level, etc. By arranging them, remote control can be facilitated. As an example of remote control command functionality, remote control system 12012 may direct any robot with illumination capabilities to activate light toward a target location, for example, to assist in optical inspection or some other visual function that would benefit from greater illumination. Input can be received from a human operator who wishes to give instructions. In this example, the remote control system receives human operator remote control commands, adapts the commands into one or more different remote control signals for robot operation orchestration units 12040 within illumination proximity of the target location, and performs proximity robot operation. Generate a corresponding remote control signal for each type of organizational unit and ensure communication of that signal (e.g., via communication management system 12010 resources) to a robot operating organizational unit to be remotely controlled by a human operator. can do. Additionally, robot operating units that receive remote control instructions may communicate, for example, among the sets (and/or subsets) that receive the signals to determine which robot operating unit, if any, is executing the command. By determining , you can further participate in the implementation of the command. The first robot thus contacted may be performing a time-sensitive function that would be disrupted if it were to divert its resources to provide commanded lighting. By coordinating with other robotic operational units, the first robot can continue time-sensitive functions based on response(s) from the other robotic operational units regarding executing remote control commands. In another remote control command example, a team of robotic operating units may, via remote control signals from remote control system 12012, allow a team of human inspectors to perform a reduced period of time while touring the work location in which the team is operating. It can be remotely controlled by instructing it to adjust its operation (e.g. to activate a quiet operating mode) to achieve noise pollution. In other remote control command examples, task-wide, team-wide, fleet-wide, or other resource-specific remote control commands may be issued to change logos due to acquisition of a fleet, temporary assignment of fleet resource(s), changes in fleet messaging, etc. The image presented on the display screen of the fleet resource(s) may be adjusted to reflect the.

Robot operating unit responsiveness to aggregated remote control signals (e.g., commands or sets of commands) may be based on broad fleet intelligence capabilities, knowledge, priorities, goals, etc. In general, the use of platform-based and/or robotic operating unit-based artificial intelligence capabilities supports broader independent decision-making capabilities for individual robotic operating units with greater situational significance.

In embodiments, remote control system 12012 may incorporate security features to prevent takeover, compromise, misuse, or interference with the control of a remotely controlled robotic operating unit. Resources used by remote control system 12012 (e.g., data storage resources, computing resources, remote control system state data, etc.) may be configured with security features such as encoding, decoding, packetization, etc. Additionally, the remote control system 12012 may operate without a human operator (e.g.) otherwise directly engaging with the remote control signaling, i.e., operating independently of the remote control signaling, e.g., autonomously, in cooperation with other robot operating units, etc. May include and/or support control override capabilities to safely obtain remote control of the robot.

RESOURCE PROVISIONING

In embodiments, fleet operating system 12002 may operate on a robot team, a robot fleet, a multi-purpose robot, and/or support resources (e.g., edge devices, communication devices, additive manufacturing systems (e.g., 3D printers), etc.) and a resource provisioning system 12014 that manages provisioning resources for robot operating units within a fleet, such as provisioning resources for. In embodiments, resources may include physical resources, digital resources, and/or consumable resources. Examples of physical resources include, for example, end effectors/manipulators, environmental shielding components, sensors and/or sensor systems, companion resources (e.g., drones, transport resources, etc.), hardware resources (e.g., specialized processing modules, data storage, networking modules, tethering modules, etc.), spare parts, human resources (e.g., technicians, operators, etc.), and power sources (e.g., generators, portable batteries, etc.). Non-limiting examples of digital resources may include software, operating parameters, task specific data sets, etc. Non-limiting examples of consumable resources may include fuel, sample collection containers, welding supplies, cleaning/cleaning supplies, deployable resources (e.g., flares, safety cones, drop zone safety nets, etc.), etc.

In embodiments, resource provisioning system 12014 may include fleet-specific inventory, local public use inventory, rental/fee-per-use based resource inventory, on-demand resource production systems (e.g., 3D printing of end effectors, etc.), third party You can provision physical resources from your inventory of physical resources, such as your inventory. In some embodiments, data processing system 12030 maintains an inventory database in one or more datastores 1203X. In embodiments, an inventory database stores inventory records, where each inventory record includes information about each resource (e.g., an identifier of the resource and/or type of resource), the general availability of the resource (e.g., whether it is available), , when it is available, etc.), price data related to the resource, and other relevant data. For example, for physical resources such as robotic units (e.g., SPRs, MPRs, and/or exoskeletons), hardware components, end effectors, and other physical components, inventory records include item identifiers (e.g., resources and /or a unique identifier that identifies the type of resource), the location of the physical resource, the physical state of the physical resource (e.g., condition of the physical resource, maintenance history of the physical resource, predicted condition of the resource, etc.), ownership data ( (e.g., who owns the resource, can the resource be purchased or rented, etc.), make and/or model of the physical resource, operational data (e.g., functionality, intended conditions and environment, weight limits, speed limits, etc.) , configuration data (e.g., system requirements, interface requirements, connectivity requirements), etc. In some embodiments, inventory may include resources that can be 3D printed. In such embodiments, the inventory record may additionally or alternatively include printing requirements (e.g., 3D printer capable of printing the resource, materials needed to print the resource, etc.), printing instructions defining instructions for 3D printing; and/or other relevant information. In embodiments, inventory records may provide inventory of digital resources such as software products, middleware, device drivers, libraries, data feeds, microservices, etc. In such embodiments, the inventory record may include an identifier for the digital resource, the provider of the digital resource, compatibility information associated with the digital resource, and access information (e.g., APIs, webhooks, and/or other methods to access or interface with the digital resource). Information), price information, functionality of digital resources, etc. can be displayed. As will be discussed, data processing system 12030 receives requests from resource provisioning system 12014 (or another suitable component, such as fleet configuration system 12020) to determine available inventory, inventory status, inventory price, etc. It can be configured to do so. In embodiments, resource provisioning system 12014 may query data processing system 12030 to determine the availability of a particular resource, the price of a particular resource, the location of a particular resource, the status of a particular resource, etc. Additionally or alternatively, in some embodiments, resource provisioning system 12014 (or another component, such as fleet configuration system 12020) configures the desired functionality of resources, robotic operational orchestration units (e.g., individual robots and/or The data processing system 12030 may be queried with the intended use of the fleet, the intended environment of the robot, and/or the compatibility requirements of the robot operating unit. In response, data processing system 12030 may return an inventory record resource corresponding to the request.

In embodiments, resource provisioning system 12014 may operate cooperatively with other systems in the fleet operations platform, such as a fleet configuration system, a fleet resource scheduling and utilization system, etc., to ensure that fleet resource provisioning rules are followed. Physical resources to be provisioned may also include computing resources, such as on-robot computing resources, robot operating unit-local fleet-control computing resources, cloud/third-party based computing resources, computing and other modules and chips (e.g., robot placed together with/within a robot operating organization unit), etc. In some embodiments, fleet resource provisioning rules may be defined in a governance standard library such that resource provisioning system 12014 interfaces with an intelligence layer to ensure that provisioned resources comply with the provisioning rules.

In embodiments, digital resources to be provisioned by resource provisioning system 12014 may include pushing software/firmware updates (e.g., to update a robot's onboard software), network network (e.g., task-specific robot configuration data, etc.) Can be provisioned through fleet configuration capabilities such as resource access credentials (to access resources), on-robot data storage configuration/allocation/usage data, etc. In embodiments, consumable resources to be provisioned by resource provisioning system 12014 may include use of specialized supply chain, task requestor resources (e.g., office setup tasks may include the use of task requester-supplied office materials, worker personal materials, etc. may be sourced from a wide range of sources, including task-, team-, and/or fleet-specific stockpiles). An example of task-related stockpiling includes stockpiling orange safety cones proximate to a long-term construction site that is accessed by the local robotic operations organization unit via resource provisioning system 12014. Use of provisioning system 12014 may include provisioning equipment, materials, software, data structures, etc. (e.g., customized end effectors) that are specifically created and/or sourced for a given work request. .

In embodiments, provisioning system 12014 may further operate in conjunction with contract systems, such as third party smart contract systems. In some embodiments, the task description may reference or include a smart contract that may include and/or result in the configuration of an instance of the provisioning system 12014 that conforms to the task description. As an example, provisioning system 12014 may receive smart contract terms that invoke provisioning constraints and/or instructions, such as from task configuration system 12018. Provisioning system 12014 may interpret these contract terms and thereby create a set of fleet and consumable resource provisioning constraints.

The examples described above for the provisioning system 12014 generally focus on provisioning related to task execution, but the provisioning system 12014 also provides provisioning of fleet resources such as compute resources, fleet configuration systems, intelligence layers, etc. Access and/or execution of fleet elements may be further handled. In embodiments, provisioning of specific resources may be enacted as part of a negotiation workflow for acceptance of a work request. As an example, provisioning a particular intelligence service (e.g., a fleet-level intelligence layer) may result in higher rates for task requesters than other intelligence services (e.g., only a robot-level intelligence layer deployed on a robot operating unit). It can result. As noted above and elsewhere in this application, intelligent services may drive value to the fleet and job configuration capabilities of platform 12000; Therefore, provisioning these systems as part of the task request negotiation may justify additional costs to the task requester.

In some scenarios, prioritization of platform 12000 resources, such as a fleet configuration system, may affect provisioning system functionality. If a job request only supports the use of those fleet resources (e.g., based on the price paid for the job) during off-peak times, then even if the platform 12000 resources are available, the platform 12000 resources will be Jobs may not be provisioned for some time.

In an embodiment, the fleet operating system 12002 may, among other things, plan and execute logistics to meet operational requirements, maintain robots, maintain availability of fleet resources (robot operating units, physical resources, etc.); Includes a logistics system 12015 that handles the pickup and delivery of parts (e.g., replacement parts, end effectors, supplies, etc.). In some embodiments, logistics system 12015 may be configured to identify the availability and locality of 3D printing resources to meet demand that may not otherwise be feasible through conventional logistics (e.g., truck-based) transportation. You can. In embodiments, logistics system 12015 may utilize intelligence services, such as machine learning systems and/or artificial intelligence systems, to recommend logistics plans.

A logistics plan may refer to the workflow created to result in the delivery of a set of items to a specific location. In embodiments, logistics system 12015 may generate a logistics plan that utilizes fleet resources, such as transport-type robots, for execution of the logistics plan. Non-fleet resources such as common carriers, leased for-hire over-the-road truck drivers, private delivery couriers, etc. may be utilized. The decision on which resources to use for execution of the logistics plan may be based on cost and availability of resources. For example, logistics system 12015 may determine that there are available fleet resources in the vicinity of the operation so that a third-party trucking service does not need to deliver available resources from a remote location, and in response, logistics system 12015 can choose available resources over third-party trucking solutions. In embodiments, fleet operations system 12002 may utilize a (platform-level) intelligence layer 12004 to assist with logistics planning and decision making.

In an embodiment, fleet operating system 12002 includes a maintenance management system 12026 that can be configured to schedule and perform maintenance on fleet resources, such as robotic operational units. Maintenance management system 12026 is responsible for field maintenance requirements, including scheduled maintenance of fleet resources in the field, to mitigate the impact on robot operating unit utilization due to movement from deployed work sites to repair depots. Requests can be processed. Maintenance management system 12026 may also coordinate maintenance and repair operations at a repair depot, etc. Additionally, maintenance management system 12026 may operate in conjunction with other platform systems, such as logistics system 12015, to ensure that maintenance is performed during the transportation of fleet resources, such as robotic operating units, between work sites. In embodiments, maintenance management system 12026 may include, provide access to, and/or integrate with mobile maintenance vehicles, spare parts depots, third-party maintenance service providers, etc. In embodiments, maintenance needs for fleet resources housed in storage areas, such as warehouses, remote inventory depots, etc., are determined by timing and timing of preventive maintenance activities for robots such that the robots are less likely to require maintenance during deployment. The same may be evaluated by maintenance management system 12026 for pre-scheduled maintenance.

In embodiments, maintenance management system 12026 may be configured to provide a robotic operating unit status through resource status reporting, which may be provided on a scheduled basis or in response to a query regarding robotic operating unit status by maintenance management system 12026, etc. You can monitor the status of fleet resources such as. In an embodiment, maintenance management system 12026 may include a robot operation orchestration unit that signals to the supervisor robot that it is experiencing reduced power output, and a robot operation unit that reports exposure to certain ambient conditions (e.g., excessive heat). Robot operating organization unit communications may be monitored for indications of potential service conditions, such as lack of heartbeat signal from the organization unit, robot operating organization unit to the robot health monitor resource, etc. Additionally, maintenance management system 12026 monitors information in a robot data repository that stores robot operating unit status information, activates self-test operating modes, and stores data that provides indication of robot maintenance needs. Probes may be placed within robot operation and/or supervision software that may perform maintenance management functions for the robot operating unit, such as collection, etc. Maintenance management system 12026 may also include maintenance robots that can be deployed with other robots in a team of robot operating units to perform requested tasks. A maintenance robot may be a configuration of multi-purpose robots deployed with a robotic team. This configuration may be temporal within the boundaries of team placement. Multi-purpose robots deployed to perform tasks in a work workflow can be dynamically (and optionally temporarily) reconfigured while deployed in teams to perform maintenance actions on other robots and fleet resources.

Maintenance management system 12026 provides human/operator input (e.g., a human observer may indicate that a robotic operating unit appears to be behaving abnormally), robotic process automation of maintenance activities, and scheduling. Artificial intelligence to predict maintenance instances, schedule them, and identify new opportunities to perform maintenance (such as replacing an air filter before performing a task in a dusty environment, under those conditions). A variety of platform services and capabilities for scheduling and executing maintenance, including leveraging machine learning to help analyze the performance of maintained robotic operating units for specific conditions before performing specific tasks. It can be configured to use . In embodiments, maintenance management system 12026 may receive maintenance-related input. Maintenance-related input may include a maintenance request from a robot operating unit (either to the requesting robot operating unit or to another robot operating unit, such as a companion robot operating unit). Maintenance-related inputs may be from edge devices (e.g., near-job-site and/or task-specific edge devices, such as fixed infrastructure devices, fleet resources, edge devices deployed at the job site by task requestors, etc.). May include requests for maintenance. Other candidate sources of maintenance-related input may include supervisor robot operations coordination units, human operators/observers, maintenance scheduling services, third party service providers, robot production vendors, and parts providers for scheduling maintenance. Maintenance management system 12026 may also use business rules (e.g., established by the task requestor, for the team, fleet, as determined by regulatory agencies, etc.) to determine appropriate maintenance workflows, service actions, required parts, etc. rules), associated tables, data sets, databases, and/or maintenance management libraries. In embodiments, maintenance activities may be assigned by a maintenance management system to fleet resources, such as maintenance robots, human technicians, third-party service providers, etc.

In embodiments, a deployed robotic operating unit may perform self-maintenance, such as correcting end effector motion, adjusting tensile structures to maintain a high degree of mobility, etc., among other things. It may consist of one or more maintenance protocols for: Self-maintenance may include, but is not limited to, a reduction in the ability to respond to detection of a damaged robotic operating unit feature, such as a rotation mechanism that no longer rotates continuously through 360 degrees. A deployed robotic operating unit may determine that a capability has been impaired and, optionally with the assistance of maintenance management system 12026, ensure that the impaired capability can be addressed when time permits rather than causing a delay in completion of the task. You can exchange assignments with robots. Additionally, robot operational unit intelligence (e.g., on-robot AI, etc.) may predict trade-offs in robot capabilities, for example, based on time-to-failure data for robot capabilities. If this predicted compromise time lands within the target task performance timeframe, the robot operating unit can request preemptive maintenance to be performed while the robot operating unit is en route to the work site. Maintenance management system 12026 may process this call for maintenance and coordinate maintenance resources to be available at the work site during the trip and/or when the robotic operating organization unit is expected to arrive.

In embodiments, maintenance management system 12026 may use intelligence layer 12200 (e.g., platform 12000 level intelligence) to predict when maintenance may be performed on robotic operational units and/or components. The intelligent service of the layer (12004) can be utilized. In some of these embodiments, maintenance management system 12026 may request a digital twin of a robotic operating organization unit from intelligence layer 12200. In such embodiments, the digital twin may reflect the current condition of the robotic operating unit, such that the robotic operating unit digital twin can be analyzed to determine if maintenance is required for the robotic operating unit. Additionally or alternatively, the digital twin service in intelligence layer 12200 may run one or more simulations involving the robot operating unit to predict when maintenance may be required. In some of these embodiments, the output of the digital twin of the robot operating unit is analyzed (e.g., using machine learning predictive models or neural networks) to predict if/when maintenance may be required. It can be.

In an embodiment, fleet operating system 12002 includes job organization system 12018. In an embodiment, the task organization system receives a task request, e.g., from a customer requesting the task. In embodiments, a task request may indicate a set of task request parameters. Non-limiting examples of work request parameters may include: Type of project and task (e.g., inspection task, packaging task, unloading task, loading task, delivery task, assembly task, monitoring task, digging task) , construction tasks, delivery tasks, etc.), budget, timeline, environment description (e.g. indoor/outdoor, size of the environment, communication capabilities of the environment, layout/blueprint/digital twin of the environment, etc.), location (e.g. (e.g., area, address, coordinates, etc.), and any other suitable parameters. In embodiments, task request parameters may indicate what type of robotic operational unit is needed and/or functionality. These and other task request details are described elsewhere in this application.

In an embodiment, work organization system 12018 may use work requests to define a work organization as a set of projects to be completed in the performance of the work, which may be sequenced in a task-level workflow. For each project, work organization system 12018 may define a workflow that defines a set of tasks to be performed upon completion of the project. In determining work organization, work organization system 12018 may determine projects, workflows, and tasks using a combination of technologies and resources, including: (i) projects, workflows, and/or tasks; artificial intelligence technology to define; (ii) a library that can define default configurations for different types of tasks and/or projects; (iii) robotic process automation; (iv) intelligent services (e.g. deep learning); and (v) quantum optimization.

In embodiments, quantum optimization may be enabled by quantum optimization system 12008, which may optimize task allocation across fleet resources, such as robot operating units, etc. Quantum optimization system 12008 may further optimize routing (logical, physical, and electronic) associated with robot fleets, tasks, teams, communications, logistics, etc. Additionally or alternatively, in some embodiments, robotic resources and other resources across a variety of fleet functions, including workforce diversity, energy consumption, compute capacity and utilization, infrastructure resource planning, engagement and utilization, risk management, compute storage capacity, etc. A quantum optimization system 12008 may be used to optimize the combination of .

In an embodiment, the work composition system 12018 and other fleet resources (e.g., fleet composition, platform intelligence, robot behavior, etc.) optimize tasks, workflows, and work execution plans, as well as, among other things, learn from failures. can benefit from the use of deep learning technology for. In this embodiment, the task configuration system 12018 utilizes a neural network and/or other machine learning model to determine task configuration based on a set of features including features extracted from the task request, platform 12000 A deep learning service can be requested from the level intelligence layer (12004). In these embodiments, the artificial intelligence service may be configured to learn task workflows, job configurations, etc.

In embodiments, task configuration, fleet configuration (which may include robot configuration), and/or task execution further improve fleet functionality, performance, and results through the use of local context adaptive task assignment, execution, resource routing, etc. It can be improved. This adaptive capability could be further enabled through peer-to-peer based communication (e.g., to a robotic operating unit within a team) that quickly and efficiently reveals the context of work activities.

In embodiments, artificial intelligence for automation of multi-purpose robotic task assignment and execution (e.g., robotic process automation through learning) may be implemented through a fleet management platform 12000, such as a fleet operations system 12002 and a platform intelligence layer 12004. Functioning in concert with elements of , it can, for example, learn robot assignments from human operator assignment activities. Other learnings that artificial intelligence systems can produce in the context of robot fleet configuration and behavior are based on outcome measures of success, including task completion, time to completion, cost of completion, quality of completion, ROI on resources, resource utilization, etc. can do.

These and other task configuration details, including the operational flow of task configuration system 12018, are shown and described in the relevant figures of this application.

In embodiments, the fleet operating system 12002 works to determine the configuration of fleet resources (e.g., robotic operating units, teams, etc.) to satisfy work requests from multiple concurrent and/or overlapping work requests. and a fleet and robot configuration system 12020 (also referred to as fleet configuration system 12020) that may work in conjunction with configuration system 12018. Fleet configuration system 12020 may determine fleet and robot configurations based on work requests, projects, robot tasks, budgets, timelines, availability of robots or robot types, configurability options for multipurpose robots, and/or other suitable considerations. You can. By way of example, fleet configuration may include specifying the quantity of each type of robot, which may be configured per job, project, task, or other organizational unit. In some embodiments, fleet configuration system 12020 may utilize platform 12000 level intelligence layer 12004 to determine fleet and/or multipurpose robot configurations. In some of these embodiments, the intelligence request may include a proposed work configuration and other related data (e.g., budget constraints, location, environment, etc.). In response, intelligence layer 12004 may output proposed fleet configurations (which may include multipurpose robot configurations). Additional details of the fleet configuration system 12020 are described and shown in the figures elsewhere in this application.

In embodiments, fleet operating system 12002 may include a job execution, monitoring, and reporting system 12022 (also referred to as job execution system 12022). The job execution system 12022 is responsible for the activities of platform functions such as logistics for robot and fleet resource delivery, data collection, cataloging, library management, and data processing system 12030 allocation to facilitate data processing activities for job execution. A job execution plan processed by adjusting may be received from the job configuration system 12018. In general, task execution system 12022 may use resources other than those configured by task configuration system 12018, such as compute, storage, bandwidth, etc., as may be defined and/or determined to be useful for executing the task execution plan. You can start by committing and managing your resources, including:

In embodiments, task execution system 12022 may further facilitate compliance with reporting requirements associated with task execution (e.g., task-specific, fleet-specific, compliance-related reporting, etc.). In embodiments, reporting may include collecting data (e.g., from robotic operating units, sensor systems, user devices, databases, etc.), processing data, and preparing feedback for use of the job execution data by job and fleet configuration systems, etc. may include. In embodiments, work execution system 12022 facilitates communication between maintenance management system 12026, resource provisioning system 12014, and robotics operational units, teams, fleets, etc. It may be assisted by other platform capabilities, such as communications management system 12010, to transmit, process, store, and manage data affecting task execution. These and other fleet and external resources may be responsible for determining the operational aspects of the requested task, e.g., which communication resources have Fleet Communications Management System 12010 reserved and/or assigned to the requested task, which robot operating orchestration units and which execute the task. Information to facilitate service and/or maintenance requirements for other resources used to do so, changes to resource provisioning that occur after operation of the job begins, etc. may be provided to the task execution system 12022.

In embodiments, job execution system 12022 may execute a job, for example, by identifying developing bottlenecks due to conditions during the job (e.g., traffic jams, unexpected ground conditions due to excessive rain, etc.). It can further facilitate the evaluation and modification of the work execution plan during the process.

In embodiments, job execution system 12022 may perform various data pipeline functions during execution of a job. In embodiments, data pipeline functionality may include optimizing the use of pre-configured sensor and detection packages combining sensor selection, sensing, information collection, preprocessing, routing, integration, processing, etc., among other things. In embodiments, sensor and detection packages may be activated by task execution system 12022 when usage is indicated as serving a range of monitoring/reporting activities. Other examples of data pipeline features include optimizing on-robot storage, selective sensor data filtering to reduce the impact on communication bandwidth (e.g., reducing demands on wireless network utilization), exception condition detection, and piping. Includes line adaptation/data filtering, etc.

In embodiments, task execution system 12022 may monitor robot power demands during task execution and address, if necessary. In this embodiment, work execution system 12022 may provide battery charge capacity (or fuel level) can be guaranteed. In embodiments, robot power demand management may include routing fleets, teams, and individual robot operating units to complete tasks while reducing overall productivity delays with integrated robot charging activities. Additional details of the functionality and operation of task execution system 12022 are described throughout this disclosure.

In embodiments, fleet functionality may be combined with 3D printing services and systems, including during job execution, to provide optional automated robotic 3D printing and point-of-use (e.g., shop floor, logistics floor, warehouse, etc.). The deployment and use of production capacity in close proximity (transportation vehicles, etc.) can enable, for example, agile, remote and flexible manufacturing as needed. Another exemplary use of fleet robot functionality with 3D printing combines these agile, flexible production capabilities with customizable product delivery for last-mile customization of products. Several example embodiments of 3D printing functionality in combination with methods and systems of fleet management are described elsewhere in this application, including, without limitation, on-robotic 3D printing of service items at a service site; 3D printing of task-specific end effectors and/or adapters based on context acquired from the shop floor; Robotic control of transportable (e.g., shop-floor-deployed) 3D printing systems; Includes 3D scanning and on-site printing.

In embodiments, task execution system 12022 may execute, deploy, and/or interface with a set of smart contracts that monitor and report to robot operations orchestration unit 12040. In embodiments, a robust distributed data system, such as a distributed ledger (e.g., a public or private blockchain), may be used to track and enhance the activities of robot fleets and/or multi-purpose robots, as well as to track and enhance related activities such as task requesters, fleet users, etc. It can be used to allocate robot resource utilization costs to parties. In some of these embodiments, distributed ledger nodes store and execute smart contracts. In embodiments, smart contracts may be configured to monitor task requests, task execution, resource usage, etc. For example, in some embodiments, the robotic operating unit may provide the smart contract with evidence of completion of the task so that the smart contract can trigger an action (e.g., pay, log, etc.) in response to the completed task. It can be configured. In another example, a robotic operating unit may be configured to report location data, sensor data, status data (e.g., charge level, component status, etc.), and/or other suitable data, whereby a smart contract may receive It can be configured to trigger specific actions based on the received data.

In embodiments, the fleet operating system 12002 may provide, among other things, access to scalable computing power for any fleet operations and/or intelligence resources, data management capabilities (e.g., data caching, storage allocation and management, etc.) , a data processing system 12030 that can provide access and control to task-related data stores such as fleets and/or libraries, fleet resource inventory control and management data structures, etc.

In an embodiment, the fleet operating system 12002 allows a user to access the fleet management platform 100 and/or (e.g., for remote control) from a remote device (e.g., a user device, VR device, AR device, etc.). It may include a human interface system 12024 that provides a human interface that allows access to individual robot operating units. In embodiments, human interface system 12024 may be configured to enter work requests (including any task-related parameters), manage fleet operations, manage fleet resources, manage fleet computing systems, software and data structures (e.g., system upgrades, etc.). ), human access to a robotic operational unit (e.g., for remote control of a robotic operational unit), augmented and/or virtual reality visualization of fleet operations, (e.g., a smart contract associated with one or more work requests) Facilitates data extraction (for creation and/or verification, etc.). As an example of use of human interface system 12024, a task requester may access status updates of the requested task via human interface system 12024. The task requester may use a remote device to observe the robot operating unit performing work on the requested task. In this example, human interface system 12024 interacts with other fleet components, such as work execution system 12022, to direct image capture resources (e.g., camera-based overhead drones) to be assigned to work tasks and to determine the current An image of a functioning robotic operating unit may be provided.

In embodiments, fleet operating system 12002 may provide support for satisfying work requests. For example, components of the fleet operating system 12002 may configure fleet resources (e.g., robotic operating units, physical modules, and/or support devices) in an efficient manner to meet work request needs, such as the timing of task execution. This can facilitate resource provisioning and logistics to ensure they are delivered to the job site. For example, in some embodiments, fleet operating system 12002 facilitates the delivery and/or maintenance tasks of fleet resources to ensure that fleet resources are allocated in an efficient manner without significantly impacting completion time. You can use a “just-in-time” strategy. In some of these embodiments, fleet operations system 12002 may be configured to anticipate fleet resource demands in response to various task requests and/or predict task execution plans that anticipate fleet resource demands and deliver and/or maintain such fleet resources. Intelligence services can be used to prepare repairs.

In some embodiments, job workflows that include multiple dependent stages can be pipelined such that no particular resource is required until other workflow stages are completed. In such a scenario, fleet operating system 12002 may delay provisioning of certain resources until previous workflow stages are nearly complete. In this way, those resources can be used in conjunction with other tasks (or other parts of the same task) while previous workflow stages are completed. In these embodiments, job execution system 12022 may monitor the status of specific tasks across multiple jobs to determine when specific resources are needed. In this embodiment, task execution system 12022 may utilize platform 12000 intelligence layer 12004 to predict when a task will be completed. In response, resource provisioning system 12014 and logistics system 12015 may operate in combination to provision and deliver resources to the work site before previous tasks are completed.

In embodiments, task execution system 12022 may anticipate task-related resource demands in a task-specific manner to predict when certain resources will be required for a particular task. For example, work execution system 12022 (operating in combination with an intelligence layer) may generate a schedule of ongoing and/or upcoming tasks for a particular work request and, in response, determine the likelihood that a particular fleet resource will be required. You can determine when it is and/or is likely to become available. Additionally or alternatively, task execution system 12022 may estimate task-related resources for a particular task in any other suitable manner. For example, a forecast of resource demand may be determined based on patterns of fleet resource demand as derived from: Work request history of the work requester (e.g., site cleanup work requests are typically requested at a work site). following completion of the task); The task requester's resource usage history from the previous N tasks performed for the task requester; Timing of work requests (e.g., requests from requestors are received on Thursday for work that typically starts the following Monday), etc. The similarity of a task requester to other task requesters (e.g., affiliated entities, direct competitors, similar SIC codes, etc.) may also form the basis for fleet resource prediction/forecast. Business relationships between entities (e.g., suppliers and shippers, sellers and buyers, consumers and recyclers, etc.) are based on actions that include work requests from suppliers/sellers/consumers, and fleet resource demands from shippers/buyers. It can form the basis for predicting timing.

In embodiments, weather forecasts and seasonal impacts (e.g., requests for snow removal and related work in northern areas during the winter season, requests for beach erosion prevention/remediation work in warm weather riparian areas on the edge of hurricane season, grassing during the spring season) Many other factors can affect fleet resource demand forecasts, such as requests for maintenance work, requests for leaf clearing work in areas with deciduous trees during the fall season, etc. Fleet resource demand forecasts also include natural disasters, vehicle accidents/emergencies, timing of social activities (e.g., assisting and incident response to stuck vehicles on high-traffic roads during rush hour, etc.), and scheduled public and/or private events. It may be activated by an event external to the core task request process, such as an event (e.g., cleaning of city streets around a sports venue after completion of a scheduled game), etc. In other examples, other sources of information that may affect projections of fleet resource needs include business objectives, such as reducing or increasing spending near the end of a financial reporting period (e.g., financial quarter, year, etc.); May include purpose. An indication that the target task requestor intends to reduce costs in the final weeks or months of the financial reporting period may indicate that fleet resources typically assigned to the task request by the target task requestor include maintenance, upgrades, pro-bono work, training opportunities, and fleet promotions. It can indicate that it will be available for other actions, such as activities, assignments to other task requesters, etc. In embodiments, fleet goals or objectives may also influence fleet resource expectations and thus corresponding readiness activities, etc. One such example is the required upgrade of a class of robots. The fleet configuration function may anticipate the need to reserve robots of that class and assign alternative robot types that can be reconfigured to meet the requirements of the reserved robot class for the duration of the upgrade activity.

In embodiments, projections of fleet resource demands may be determined through use of a fleet management platform 12000, such as platform 12000 intelligence layer 12004 and fleet operating system 12002. For example, in some embodiments, platform 12000 intelligence layer 12004 may provide weather forecasts, public activity calendars, work request data (e.g., timing, work parameters, relationships to other work requests, etc.), social media, etc. You can analyze source data that can affect fleet resource demand, such as postings, government activity/legislation, seasons, etc. In this example, the platform 12000 intelligence layer 12004, acting in conjunction with the fleet operating system 12002, may predict fleet resource demands based on analysis of disparate data sources (e.g., using neural networks, etc.). . In this embodiment, platform 12000 intelligence layer 12004 can process data from heterogeneous data sources and determine likelihood of fleet resource demand over a range of factors.

Other aspects of fleet resource forecasting include fleet preparation and/or by automatically configuring one or more work requests for fleet preparation-oriented activities (e.g., preparation and/or maintenance of a robotic operating unit or support device). or may include use of the work request process described herein for maintenance activities. In this way, fleet management platform 100 may operate to facilitate performance of work requests while ensuring that fleet-specific needs (e.g., maintenance) are met. Balancing work requests from clients of platform 12000 with fleet self-intensive activities (e.g., maintenance) and work expectations may be achieved through the use of relative weighting of work requests.

In embodiments, fleet management platform 12000 may use external data sources 12036 to perform various platform functions, including task composition, fleet configuration, task negotiation (e.g., via a smart contract facility), task execution, etc. You can interface with Examples of external data sources for use by platform 12000 include value chain entities (e.g., third parties paying for fleet services, etc.), task context to perform teaming and/or execution of requested tasks; Includes enterprise resource planning system (ERP), smart contracts, etc. that can provide. Other external data sources include third-party sensor systems (e.g. GPS data, value chain logistics data on when materials needed for a job should be delivered, etc.) as well as third-party data streams (e.g. weather, traffic, etc.). , electricity price, etc.).

In some embodiments, fleet management platform 12000 may support the use of smart contracts related to work requests, work performance, resources, allocation, etc. In embodiments, work requests may be routed through a smart contract handler that captures work requirements, requester goals and objectives, and fleet work execution constraints into a dynamic smart contract. In some embodiments, a smart contract provides negotiated routing of a multi-purpose robot from a first location (e.g., current shop floor, warehouse, temporary storage/service location) to a second location (e.g., target shop floor). It can be used across fleet management platforms to address all manner of fleet operations, such as managing As a further example, smart contracts can be deployed as controls for a bidding system for robot time/task utilization. As another example, a smart contract may monitor certain activities associated with a work request (e.g., task-related activities, etc.). Smart contracts rely on and/or derive from access to fleet platform data (e.g. task progress, sensor data, etc.) to trigger actions defined by the smart contract, such as payment upon completion of a task or task. You can make a profit. The fleet management platform 12000 may provide access to fleet resources, including fleet data, through application programming interfaces, infrastructure elements such as sensor networks, edge computing systems, etc., to update status related to smart contract terms and conditions. You can.

Referring to the embodiment shown in FIG. 134 , job organization system 12018 and fleet organization system 12020 collectively generate a work execution plan 12310, in accordance with some embodiments of the present disclosure. In an embodiment, the task execution plan 12310 may define a set of tasks to be performed upon completion of the requested task and may further define the configuration of a fleet of robot operating units that will complete the task. In embodiments, work execution plan 12310 may include task definition 12304D, workflow definition 12306D, fleet configuration 12020D (which may include robot configuration of individual robots), team assignment, and shop floor details. It may include reference to (or integration of) contextual information such as the like. In an embodiment, task configuration system 12018 receives a request 12300 defining a task to be performed, and task configuration system 12018 generates a task definition ( 12304D) can be determined. In an embodiment, task configuration system 12018 further defines a set of workflow definitions 12306D. Workflow definition 12306D defines at least one order in which tasks are performed upon completion of a project and/or task, including any loops, iterations, triggering conditions, etc. In embodiments, work organization system 12018 may determine a workflow 12306D based on a task definition 12304D that includes tasks and/or projects. Task composition system 12018 may utilize a library of pre-configured workflows to complete specific tasks. Additionally or alternatively, the task organization system 12018 may utilize the platform 12000 intelligence layer 12004 to obtain an initial workflow definition 12306D for tasks and/or projects that are part of a larger task. In some embodiments, a human may construct the initial workflow definition and/or provide input used to determine the initial workflow definition. In embodiments, work configuration system 12018 may interface with one or more components of fleet management platform 100 to exchange information and/or utilize one or more services to develop robotic fleet work execution plans 12310. there is. For example, task configuration system 12018 includes data processing system 12030, robot configuration library 12314 of robot, fleet, project, and task-related information, fleet-level intelligence layer 12004, and fleet configuration system 12020. ) can be interfaced with, etc.

In the example of FIG. 135 , work organization system 12018 may include a plurality of systems that perform work plan preparation functions by processing information received in work request 12300. In an embodiment, the systems of task composition system 12018 may include a task parsing system 12302, a task definition system 12304, a workflow definition system 12306, and a workflow simulation system 12308. In the illustrated example, the task organization system 12018 systems operate in combination to generate a task execution plan 12310 that is used to define a set of robot operating organization unit assignments 12312. In embodiments, robot operations organization unit assignment 12312 may be supplemental to or integrated with task execution plan 12310 and may identify specific robot teams and/or robots assigned to each task. For example, robot operating unit assignment 12312 may define specific tasks, and for each task, a robot unique identifier and/or a specific robot team assigned to the task with a team identifier assigned to the task. A specific robot can be identified. In embodiments, robot operations organization unit assignments 12312 may be generated by job organization system 12018 and/or fleet organization system 12020.

In an embodiment, task parsing system 12302 is used to determine a set of task request parameters that are ultimately used to determine task definition, project definition(s), task definition, workflow definition, fleet configuration, and robot configuration. Receive and parse the work request (12300). In embodiments, task parsing system 12302 may receive tasks from a user via a user interface, such as human interface system 12024, to receive input by an operator to configure, adapt, or otherwise facilitate parsing of task requests. Requests can be received. Additionally or alternatively, task parsing system 12302 may receive task requests from client devices associated with the requesting organization.

In embodiments, task parsing system 12302 may be a collection facility for receiving electronic versions of task descriptions and related documents, such as drawings, material lists, flow charts, GPS data, smart contract data and/or terms, links thereto, etc. It can be composed of: The collection facility may parse the document for keywords, references to activities, etc., which may be particularly useful in determining which aspects of the described task may be suitable for a robotic task. In an example, the collected documents may have structural and/or content elements (e.g., numbered indentations) that may facilitate identification of tasks, subtasks, sequences of tasks, dependent requirements for tasks, workflow descriptions, etc. Lists, references to robot identifiers, references to existing robot task content, etc.) can be processed with content and structural filters to detect those parts for robotic automation. Additional keywords within the collected task content, such as weight terms, task environment terms, etc., can be usefully applied by the task configuration system 12018 element by providing insight into the type(s) and configuration of robots required. there is. As an example, a keyword suggesting an object to be moved would suggest a robotic transport device/team with a weight of 14 tons and a movement capacity of at least that amount.

In embodiments, task parsing system 12302 may be used to improve techniques for parsing task content, which may include descriptive data (e.g., as may be provided by platform 12000 intelligence layer 12004). (such as) may incorporate and/or utilize machine learning functionality. In addition to machine-based learning through human-generated feedback on task content parsing results, learning can be done on common and specialized knowledge bases such as other task content parsing tasks (e.g., previous task requests), technical dictionaries, and experts. It can be based on experience.

In embodiments, task parsing of task content may include automated parsing of structured and unstructured text. In some embodiments, task parsing system 12302 may be configured to identify (and optionally resolve) missing/unclear data and eligible task content data (collectively referred to as “insufficient information”). In response to identifying insufficient information, task parsing system 12018 may generate and provide a request to a human operator via a user interface for clarification of the insufficient information. These requests can identify specific input from the user to provide, so that the requests identify disambiguating content that was initially missing or unclear. Additionally or alternatively, parsing system 12302 may determine disambiguation content (e.g., via a query) from a library 12314 that maintains data from previous operation requests, and thus Disambiguation content can be obtained using information and context. If the parsing operation cannot determine disambiguation content, parsing system 12302 may generate a request to disambiguate content as discussed above.

In embodiments, the scope of job description information may be provided to, determined, and/or extracted from job organization system 12018. Examples of task request parameters may include, but are not limited to: (i) physical location information that may be used to automatically determine transportation options, operating restrictions, permits, movement restrictions, local assets, logistics, etc.; (ii) available field power voltages, frequencies, currents, etc. may limit available equipment or require additional equipment, especially for support; (iii) Digital data on site layout, such as 3D CAD models, scans, and robot surveys, can be available or completed as part of the initial project scoping, determining task prioritization and workflow routing, robot selection, supervision needs, etc. Can be used to automatically provide; (iv) operating environment, including temperature, hazard statement(s), terrain, weather, etc.; (v) deliverables such as data, reporting, analysis, etc.; (vi) customer interfaces for data exchange, such as network interfaces, APIs, and security; (vii) availability of telecommunication networks such as landlines, 4G, 5G, WiFi, private networks, satellites, connectivity constraints, etc.; (viii) budgetary constraints on equipment limitations, field time, and permits; (ix) Scheduling for field availability, reconfiguration flexibility, earliest start time, latest finish time, activity rate such as number of robots active at any given time, etc. Examples of other job description information that can be handled by a job parsing system include smart contract duration, certification level of robot motion software for robots deployed on the job site, insurance period, site access requirements (e.g. can only be accessed when no human is present or through coordination with a human present on-site), and may contain contract-related information, such as conditions for assigning proxies for tasks, activities, workflows, or overall tasks. there is.

In embodiments, task organization system 12018 (e.g., task parsing system 12302, task definition system 12304, workflow definition system 12306) may combine robotic automation task content with other task content (e.g., , cost, payment, financing, etc.), a library (12314) to identify pre-configured candidate tasks, workflows, and/or content and structure filters to distinguish them from complete task configurations that substantially meet the requirements of the work request. ) can be referred to. In embodiments, library 12314 or another task configuration library may facilitate mapping representations of task content with target terms representing robotic automation. As an example use of automated tasks from library 12314, the requested data collection task may include a requirement to sample surface water from a storm system catch basin. Task parsing system 12302 may identify sampling requirements, and in response, task definition system 12304 may generate a library that meets the requirements of that portion of the task request description that can be used to define a task execution plan 12310. (12314) An automated sampling task may be identified for sampling water within. If the task configuration system 12018 determines that a suitable task configuration is available (e.g., from library 12314), such as if the requested task has previously been requested, the task configuration system 12018 The previous task execution plan 12310 corresponding to the requested task can be used as a proposed task execution plan 12310 for further verification by the current fleet standard, etc. For example, the platform intelligence layer 12004 may be directed to a set of governance standards (e.g., with one or more intelligence services, including without limitation machine learning services) to ensure that proposed work configurations comply with said standards. The proposed work configuration can be analyzed. Platform intelligence layer 12004 may perform other intelligence-based tasks for the proposed task configuration.

In some scenarios, task organization system 12018 may determine that one or more tasks, workflows, routines, etc. do not have a suitable counterpart in library 12314. In this scenario, the task parsing system 12302 determines the robot-fleet focus requirements (e.g., task definition parameters, robot configuration parameters, proposed task sequences, etc.) to perform the task, which are then passed to other task configuration system modules for processing. ) can be created. In embodiments, task parsing system 12302 may rely on platform 12000 intelligence layer 12004 for suggestions of task requirements, including combinations of tasks that, when selectively adapted, can meet such requirements. . In an example, operational requirements may include sampling surface water from a frozen storm catch basin. In this example, library 12314 may not include a frozen surface water sampling task. However, platform 12000 intelligence layer 12004 may recommend an ice melt task followed by a water sampling task to meet operational requirements.

In embodiments, task parsing system 12302 may include and/or interface with analysis modules/governance libraries of intelligence layer 12004 of platform 12000. Task parsing system 12302 may utilize governance-based analysis by providing candidate robotic automation portions of task content (e.g., terms, etc.) for processing. Intelligence layer 12004, in response to the provided portion of work content, provides and/or displays one or more of the safety standards and/or one or more of the operating standards to be applied during preparation of the work execution plan by work organization system 120118. can do.

In embodiments, task parsing system 12302 allows task composition system 120118 to define robotic tasks, configure fleet resources, define workflows, simulate workflows, generate task execution plans, etc. It can include a task requirements module, which creates a set of task request instance-specific requirements for use when performing a task. In embodiments, a set of task request instance-specific requirements may be determined based on at least one or more of the following: (i) candidate portions of task content indicative of robotic automation (e.g., terms indicative of robotic tasks); (ii) one or more inputs from a user interface (e.g., disambiguation of terms), (iii) safety and operational standards (e.g., from a governance layer), and (iv) recommended robot tasks and associated context information. (e.g. provided by the fleet intelligence layer).

In embodiments, work content parsing system 12302 may use content filters and/or structural elements to identify structural elements within work content that may represent one or more of tasks, sub-tasks, task sequencing, task dependencies, task requirements, etc. Filters can be applied. In embodiments, the detected structural elements may facilitate selection and configuration of robotic operational organizing units, for example, by fleet configuration system 12020. In an example, a structural element that distinguishes a set of tasks may be used by a fleet configuration system to avoid assigning the same robot operating unit to tasks within the set of tasks and tasks outside the set indicated by the structural element.

In embodiments, task parsing system 12302 may integrate and/or utilize a task request composition agent/expert system that can be configured to facilitate developing task description parsing capabilities.

In embodiments, task definition system 12304 may organize task data into task definitions 12304D (e.g., discrete robotic tasks or tasks performed by a robotic team). Task definition system 12304 may further coordinate other systems in task composition system 12018, such as workflow simulation system 12308, to optimize task definition.

In embodiments, task definition system 12304 may compile task data compiled by task parsing system 12302 to facilitate defining the individual actions of one or more robot operating units within a fleet of robots in the performance of requested tasks. can be improved. Defining a task may be based on information about robots capable of performing the defined task, robot type, robot characteristics, and robot configuration. In an embodiment, task definition system 12304 may facilitate task definition 12304D to facilitate fleet configuration system 12020 in determining the use of general/multipurpose robots, special purpose robots, and/or a combination for each defined task. ) information can be provided additionally. In an embodiment, task definition system 12304 may define a task that satisfies a first flit object of a set of flit goals. The first fleet object may include defining tasks that can be performed by a multi-purpose robot, for example, by decomposing the work content into smaller tasks that require less customization of the robot. In embodiments, task definition system 12304 may reference libraries 12314, platform 12000 intelligence layer 12004, or other platform-specific or accessible resources when performing task suggestions.

As task definition system 12304D defines the tasks of a job, the task definitions may be cataloged and stored for future use, for example, in library 12314. In some embodiments, task definition system 12304D may adapt task definitions from previously cataloged task definitions (e.g., tasks for a specific type of environment or specific conditions from a previously cataloged task definition). Adapting the definition). In this embodiment, task definition system 12304D may catalog derived task definitions in library 12314 along with adaptation instructions. In some embodiments, a task definition that is cataloged in library 12314 may be associated with an already cataloged task definition and/or may replace an already cataloged task definition and be cataloged as a sub-task of an existing task. It can be done, etc. In general, a task definition may include associated tasks, serialized tasks, nested tasks, etc.

Information about jobs may be stored in library 12314 for future use, so task definition system 12304 can access library 12314 to retrieve information about jobs, robots, fleets, etc. In the present example embodiment of inspection of a ventilation system, information accessible through library 12314 may include, for example, methods for accessing information regarding the physical configuration of a ventilation system. Task definition system 12304 may also be used to update information, such as by adding one or more tasks to a list of tasks for a ventilation inspection task, resulting from optimization of task definitions performed by the task execution system, etc. Library 12314 is accessible.

Optimization features of the task definition system are described below in relation to feedback from other elements of the task composition system 12018, such as the workflow simulation system 12308.

Task definitions may be created and provided to other elements of the work composition system 12018, such as the workflow definition system 12306 and the fleet composition system proxy 12305. In embodiments, a fleet may provide task definitions (and other suitable information) to the fleet configuration system 12020. In an example, the fleet configuration system proxy 12305 may configure the fleet configuration (e.g., based on geography, timing, etc.) As indicated), the set of candidate robots to perform the task can be narrowed down to a specific robot type (and optionally to a specific robot within the fleet). Fleet configuration system proxy 12305 may process task definitions, which may include robot identification information (e.g., robot type, etc.), to align the fleet's resources with relevant task information. In one example, the fleet configuration proxy 12305 may use the fleet resource provisioning system 12014 to perform fleet resource allocation, scheduling, etc. in support of at least some of the goals of a work request being processed through the work configuration system 12018. Data suitable for use by fleet operation elements such as can be generated. Fleet configuration proxy 12305 may use fleet configuration modeling to determine candidate fleet configurations that meet job requirements. Modeling can be useful in determining impacts to fleet resources that can be considered during fleet configuration functions, resource allocation, etc. In embodiments, fleet configuration modeling may include the use of platform intelligence layer resources, such as machine learning, artificial intelligence, etc., in determining one or more preferred fleet configurations that also meet one or more job description requirements. Fleet configuration system 12020 is described in more detail elsewhere in this disclosure.

WORKFLOW DEFINITION SYSTEM

In an embodiment, work organization system 12018 receives task definitions from task definition system 12304, receives fleet configuration information from fleet organization system 12020, and organizes tasks (e.g., deliverables and/or tasks and a task workflow definition system 12306 that receives other work request information that may facilitate the timing of the task workflow and generate one or more task workflows based thereon. In embodiments, workflow definition system 12306 uses task definitions, task parsing systems, and output from real-time external data, such as maintenance management systems, ERP systems, etc., to identify workflow possibilities to determine task workflows. Contains information from the fleet management system for In embodiments, a task workflow defines the order and manner in which tasks are performed to perform a project/task. In embodiments, workflow definition system 12306 may apply job description information to a set of task definitions and fleet configuration data to create one or more workflows for performing one or more activities of a job. By way of example, a workflow may cover activities such as entering a ventilation duct through a portal such as a ventilation inlet port, etc. The tasks defined for this activity can be collected into a workflow or part of it, ordered to ensure proper compliance with task requirements, and published as a set of requirements for performing the activity/workflow. The task workflow definition includes information describing the quantity and type of robots, tools/end effectors, etc. that may be provided by the fleet configuration system 12020 for one or more tasks sequenced by the workflow definition system 12306. can do. In embodiments, this portion of the workflow definition may be used to identify and determine the required configuration of, for example, one or more robots to be prepared prior to performing tasks in the workflow (e.g., defined in the workflow). other modules of the task configuration system 12018 (e.g., task execution system 12022) to ensure that the multi-purpose robot is (re)configured into a configuration that enables it to perform the task before performing the task. It can be used by . Other information generated from the task execution plan may include a sequence of tasks (e.g., as generated by a workflow system), which may further identify the sequence of robots required to perform the task. there is.

The workflow definition system may utilize the resources of the robot configuration library 12314 when defining the workflow. Workflow definition parameters, such as how to determine minimum time between tasks, coordination between tasks, task classification, workflow scope, etc., may be available in library 12314 and/or in information retrieved from work requests. These and other parameters can be set to default values, but may also include task-specific variables that can be adjusted, for example, by a workflow definition system to meet task-specific needs. An example of using robot configuration library 12314 information to develop a task workflow definition may include a robot movement task followed by a sampling operation. Information in the robot configuration library 12314 related to the material/object to be sampled may indicate, for example, that a minimum residence time after the robot is deployed must be satisfied prior to sampling, to allow ambient dust to settle, etc. . Other useful information that the workflow definition system can utilize from the robot configuration library 12314 may include templates, pre-configured or default workflows, such as workflows developed for previous execution of the task. The workflow definition system may determine which workflows (base workflows) within library 12314, if any, are suitable for use in the current task workflow definition instance; determine adjustments to discovered workflows; It is configured to create an instance-specific task workflow, etc., which may include additional tasks not found in the base workflow and/or exclude unnecessary tasks found in the base workflow.

Other examples of robot configuration library 12314 information that may be useful in developing task workflow definitions include the availability of sensor detection packages. These sensor detection packages can indicate a preferred sequence for detecting tasks and thus influence the workflow of those tasks. These and related reconfigured sensor and detection packages can combine sensor selection, sensing, information collection, preprocessing, routing, integration, processing, etc. These sensor and detection packages may be included in the fleet configuration process, such as being included in a job execution plan for use by job execution, monitoring, and reporting system 12022. In embodiments, its use is indicated as serving a range of monitoring activities, etc.

A job workflow definition system may examine task-to-task dependencies to identify potential workflow independence and dependencies, especially when constructing a job execution plan that may involve parallelized use of fleet resources such as teams (e.g. For example, performing a second task depends on completing the first task).

Features of the intelligence layer, such as team twin capabilities, fleet twin capabilities, etc., may also be beneficially applied to simulate and verify workflows, such as in the workflow simulation system 12308 of the work configuration system 12018. Workflow simulation system 12308 may perform a simulation of portions of a work configuration, such as portions organized into a work workflow by a workflow definition system. In an example of a workflow simulation, a set of tasks defined by a task definition system and organized as part of a job workflow can be divided into robots, tasks, workflows, etc., such as robot twins, task twins, workflow twins, team twins, and fleet twins. can be modeled using functional equivalents for These twins may be retrieved from library 12314 and executed by the processor to simulate a set of tasks, such as verifying a defined task. In embodiments, the fleet intelligence system may, for example, apply workflow definitions and task definitions to one or more workflow models and/or task/robot/fleet twins operating in an artificial intelligence environment machine learning environment, thereby creating a simulation of such workflows. It can be used to provide at least some of it.

Workflow simulation system 12308 may also generate feedback from simulating workflows defined by the workflow definition system, which may be useful for improving workflow definition, task definition, robot selection, etc.

Workflow simulation system 12308 may establish or otherwise access criteria to determine whether the workflow meets criteria such as completing tasks, operations, etc. in a timely and successful manner. By applying these criteria for measuring the results of the workflow simulation, the workflow simulation system 12308 can determine one or more workflow options, workflows, etc., before providing feedback to, for example, a task definition system, task parsing system, etc. Robot options, fleet configuration options, etc. delivered according to the definition system can be verified. Options that do not meet the criteria (e.g. consume excessive resources, cause wear and tear on the robot, do not meet schedule, etc.) are considered to improve task organization features, such as structuring tasks into workflows, etc. It can be marked like this.

Additionally, workflow simulation system 12308 may utilize the platform 12000 intelligence layer. In embodiments, the platform 12000 intelligence layer may provide access to and behavior of instances of fleet twin modules that may provide critical understanding of fleet-based impacts on workflow definitions to perform requested tasks. You can. In embodiments, a logistics twin of a fleet intelligence system may provide useful workflow simulation information through the modeling of the delivery and costs of robots, personnel, support equipment, etc. for the delivery of a fleet of robots to a work site. This modeling of fleet logistics allows local fleets that will soon become available (perhaps after the preferred start date of the requested work) to do the work at a lower cost than using currently available crews who will need to transport the logic and transport to the job site. It can be revealed that it can be completed. In embodiments, fleet twins may facilitate identifying available robot operating assets during scheduled work by modeling fleet behavior, such as robot maintenance requirements for robots during preferred task execution times. In embodiments, the task twin capabilities of the Fleet Intelligence System may enable multi-purpose robots to perform different tasks, e.g., (i) moving a ventilation inspection wand to a ventilation system port; (ii) collecting debris being removed from the ventilation system. and movement) can facilitate modeling of the robot configuration when reconfigured during a task (e.g., during a task). The task twin capabilities of fleet intelligence systems can optionally benefit additional workflow definition clarity through workflow simulation by applying a virtual set of pre-configured robot twins to perform the candidate workflow or portion thereof that is being defined. . In embodiments, the team twin capabilities of the fleet intelligence system may be implemented by the workflow simulation system of the work configuration system 12018, for example, by using a team of pre-configured robots to operate and validate candidate workflows prepared by the workflow definition system. can benefit.

In embodiments, the results of a workflow simulation may include one or more data structures suitable for use in task execution planning.

In addition to task definitions, robot definitions, workflow definitions, fleet configuration parameters, etc., work execution plans can be configured with smart contracts, work resources (e.g. You can identify contracts for work, such as delivery times for fleets, schedules for deliverables, etc.

In an embodiment, fleet configuration system 12020 configures the fleet's resources for work based on task definitions and/or workflow definitions. Fleet configuration system 12020 may determine fleet configuration based on other considerations such as budget, environmental conditions, time constraints, available inventory of robots and/or parts, etc. Fleet organization system 12020 may operate in conjunction with work organization system 12018, such as when tasks are organized into workflows. The task definition may define, for example, a task that can be performed by a special-purpose or multi-purpose robot. Because job workflows may be affected by the availability of each type of robot, job organization system 12018 may utilize fleet organization system 12020 in determining candidate job workflows. As an example, a workflow involving allocation by a fleet configuration system of special purpose robots (e.g., a special purpose robot may be provided for the task for which it is configured) may be modified to take into account differences between these types of robots ( may need to be adjusted (compared to workflows utilizing multi-purpose robots). A special purpose robot can perform a task or tasks more efficiently and/or with greater precision than a general purpose robot; Accordingly, special purpose robot workflows can be configured with shorter completion times (e.g., greater robot efficiency) or without independent verification steps (e.g., greater precision or self-verifying special purpose robot capabilities). These are just examples to illustrate the potential for impact on the workflow definition of a fleet composition system.

In embodiments, configuring a fleet for a requested task may include organizing fleet resources into teams of robots assigned to specific tasks and/or projects (a robot or team of robots may be assigned multiple tasks and/or projects). Please note that you may receive it). Each robotic team may include one or more robotic operational units, which may include any one or more of special purpose robots, multipurpose robots, rigid and/or soft robots, exoskeleton robots, humans, working animals, etc. Additionally, the constructed robot teams may be task specific, and team membership may be temporary for any given robot operating organizational unit. By way of example, a special purpose welding robot configured to perform a welding task, or optionally a multi-purpose robot, may be assigned to a first robotic team only for the duration of time that the welding task is being performed by the first robotic team. The same welding capable robot may also be assigned to a second robot team only for the duration of time that the second robot team welding is being performed. Time sharing of fleet resources, such as welding-capable robots, can be communicated, for example, from fleet configuration system 12020 to a job configuration system, such that the workflow defined by the job configuration system is such that the workflow of the welding-capable robot for each robot team is Availability can be considered. In embodiments, any given robot or group of robots may be assigned to multiple teams spread across multiple tasks by fleet configuration system 12020 using a robot-specific time sharing approach or other resource utilization optimization techniques. there is. In one example, fleet organization system 12020 may use a multidimensional robot utilization planning system to assign each robot within a fleet to one task for a time unit, such as a day, hour, or fraction thereof, such that each robot in the task organization system Allows instances of to request use of the robot for a specific time (e.g., Tuesday the 23rd from 10:00 AM to 4:00 PM) or an amount of time units (e.g., 6 consecutive hours). Fleet configuration system 12020 may respond to the request with a robot fleet configuration description notifying task workflow definitions, etc.

In embodiments, fleet configuration may be a multipurpose robot (e.g., as may be indicated by a task definition system, etc.) to configure multipurpose robots to be included in a team or fleet of robots for the performance of one or more tasks in a task. Additional configuration information may be included. Multipurpose robot configuration information may define modules that can be combined with the robot, including end effectors, synchronous adapters, sensors, image processing modules, special purpose processing modules, communication modules, etc. Multi-purpose robotic modules and applications are further described elsewhere in this application.

In some embodiments, configuring a fleet for a requested task may include allocating robotic support resources such as edge devices, charging capabilities, local data storage capabilities, shipping containers, docking stations, spare parts, required technicians, etc. In embodiments, a fleet configuration system may also assign robots to distinct roles, such as roles related to team organization (e.g., supervisor), security, human interaction, inspection/quality control, etc. These roles may not be individually defined in the work request; However, criteria in work requests (e.g., quality inspection reports) may lead to the assignment of these robot roles. In embodiments, fleet configuration system 12020 may assign some team roles to humans, including human team member participation requirements, support, equipment, etc. Fleet composition systems can consider human safety when assigning humans as team members. As an example, human team members may be required to wear safety face shields when participating in a team that performs welding operations.

In embodiments, fleet configuration system 12020 may utilize a library to determine fleet configuration. In this embodiment, fleet configuration system 12020 may determine team composition for a defined task or project using a library 12314 that defines different configurations to perform a particular task, thereby using a lookup table or other Associations are used to determine team composition for a given set of tasks. In embodiments, library 12314 may include properties for different robot types, such as multipurpose robots. As an example, an attribute of a multipurpose robot may indicate the minimum size of the multipurpose robot. In embodiments, fleet configuration system 12020 determines the types of robots that can perform a task based on one or more task request parameters and attributes identified by task parsing system 12302 (and optionally configured as a task definition). You can filter. A task or work movement can be performed in a space smaller than the smallest available multipurpose robot (e.g., based on data generated by the task parsing system 12302, existing task execution plan 12310, task request 12300, etc.). When requesting access to, fleet configuration system 12020 will not include a multipurpose robot; Instead, it will attempt to identify different robots and/or robot types/configurations that can meet the size requirements. In embodiments, fleet configuration system 12020 may reference combinations of robot sizes/types, etc. to fit the requirements of a defined task. Additionally, fleet composition system 12020 may propose two robots to perform a task when one may not meet the other requirements of the task. In a simple example, a task that involves moving long distances and then performing actions in a small space may involve a multi-purpose robot that moves long distances efficiently (and optionally includes a payload-carrying capability suitable for transporting special-purpose robots) and This can be solved by a fleet configuration system with a combination of robots such as special purpose robots that meet small space requirements. In embodiments, fleet configuration system 12020 may pass a fleet definition including multiple robots, robot types, robot configurations, etc. to task configuration system 12018. A general goal of the fleet configuration system 12020 may include creating a fleet configuration(s) that requires the fewest number of robots and/or robot types for proper execution of some of the requested tasks. However, fleet configuration system 12020 may operate in cooperation with task definition system 12304 to create a task-specific fleet configuration that includes more than one robot type/configuration/combination, thereby allowing fleet management system 12000 ) allows other elements of the task to efficiently manage the execution of the requested task. This fleet configuration may include robots, etc., that other elements of the work organization system (e.g., a work workflow generation system) may consider, for example, when organizing a plurality of defined robot tasks into a work workflow 12306D. A preferred robot and/or combination of robots may be indicated to meet a goal, such as efficient use. Accordingly, the fleet configuration may include first, second, and third robot representations to perform the task. Alternatively, a fleet configuration for a work request may identify multiple robots, each robot being assigned a utilization weight based on criteria such as efficient task completion, profitability, preference for using fleet robots, etc.

In embodiments, fleet configuration system 12020 may include robots and/or modules (e.g., physical modules and/or software modules) available to configure a multipurpose robot, the locations of those robots and/or parts, and the status of the parts. Inventory data repositories can be consulted to determine, for example, whether maintenance is scheduled or required for available robots or parts. In this way, fleet composition for jobs, tasks, teams, etc. may be determined by available inventory of robots, modules, support equipment, and/or spare parts. Additionally, a fleet maintenance management system as described in this application may be used to determine which robots are held in reserve from use for critical maintenance, which robots may be deployed, but which are not required for service and/or maintenance or other concerns; Status of the robot that can be added to and/or replenished in the inventory data store, such as the status of spare parts or whether it has reduced capabilities due to other service activities (e.g. deadline, current location, expected installation, etc.). The pattern can be traced. Accordingly, fleet configuration system 12020 may consult and/or be notified by the fleet maintenance management system regarding fleet resource maintenance knowledge that may impact operations. Additionally or alternatively, the fleet configuration system 12020 may request fleet configuration from the platform 12000 intelligence layer 12004, where the artificial intelligence service 12028 includes task definitions, workflow definitions, budgets, It may receive as input a set of parameters including environment definitions, task timelines, etc., evaluate a plurality of candidate fleet configurations, and determine a target fleet configuration capable of performing the task. In embodiments, humans may define or redefine any part of the fleet configuration through the human interface of the fleet configuration system.

In embodiments, job and fleet configurations may be fed to a digital twin system, whereby the digital twin system can perform a simulation of the job given the job and fleet configuration. Job configuration system 12018 and/or fleet configuration system 12020 may be configured to optimize (or substantially optimize) one or more parameters, such as job timeline, overall cost, robot downtime, maintenance-related downtime, delivery cost, etc. Job configuration and fleet configuration can be redefined iteratively. Once the work configuration system 12018 and the fleet configuration system 12020 determine the fleet configuration, including task and workflow definitions, as well as multipurpose robot configuration and team assignment, the fleet management platform executes tasks corresponding to the work requests. The plan (12310) can be printed.

In embodiments, fleet configuration system 12020 may utilize digital twins when configuring fleet resources. Use of digital twins with fleet configuration may include identifying and/or defining one or more digital twins of one or more robots based on information in task definition 12304D. Fleet configuration may include identifying the configuration and/or operation of the multipurpose robots so that the multipurpose robots can perform the task or portion thereof. These general-purpose (and optionally special-purpose) robot task configuration instructions may be generated through the use of a digital twin for one or more of a set of candidate robots to perform the task. In an illustrative example, a multipurpose robot may be associated with a plurality of configuration/operational data structures for configuring the multipurpose robot to perform routines, actions, tasks, etc. Fleet configuration system 12020 may identify or otherwise be provided with one or more candidate multipurpose robot configuration data structures (e.g., from library 12314) for use to perform a task. Some of these candidate configuration data structures may include rotation rates for the end effector to secure the panel rotation retaining bolts. The requested operation requirement may explicitly or implicitly indicate that the turnover rate for securing the panel is different from the value in the candidate configuration data structure. In an embodiment, a fleet configuration system makes any adjustments to a candidate configuration data structure (e.g., reducing turnover) and applies them to an instantiation of a digital twin of a candidate multipurpose robot with the adjusted configuration data structure. The execution (e.g., simulation) of the digital twin may be observed and/or evaluated, and stored in a library 12314, etc. Newly stored configuration data structures may be cataloged for efficient access in the future based on the job request and/or other parameters of the requested job, task, etc.

Robot Configuration Library 12314 contains job information, robot information, fleet information, task definition rules/metadata, workflow configuration rules and/or techniques, and job configuration systems that may be useful in determining how to define robot tasks. Results of previous work requests from applications (e.g., previous work execution plans), etc. This library 12314 may be accessed and/or updated by the functionality of the operational platform. Illustrative examples of library 12314 are described variously in this application along with task operation platform functions and features such as task configuration and the like. As an example, robot configuration library 12314 may contain specific references to configurations of multipurpose robots that may be used during fleet configuration, task execution, etc. In this example, robot configuration library 12314 will have a reference to a set of robot configuration data (e.g., data that, when uploaded to a multipurpose robot, may enable the robot to perform functions such as standing, welding, etc.) You can. Additionally, the library may provide cross-references of multipurpose robot configurations with other robot-related information, such as base model, version, required features, etc., that may be required for successful deployment of a robot configured with a given configuration. Additionally, the library may suggest alternatives for specific combinations of robots and configurations, such as indicating that newer versions of the robot model may include built-in capabilities provided by specific configurations. Accordingly, the fleet configuration system can have greater flexibility in deciding which robots to deploy for different tasks. References to library 12314 are made in this application using context modifiers such as robot configuration libraries, etc. These context modifiers may suggest one or more portions and/or instances of library 12314 for illustrative purposes only.

In embodiments, optimization features of the task definition system are described below in conjunction with feedback from other elements of the task composition system 12018, such as the workflow simulation system 12308.

Figure 136 presents a flow diagram illustrating an embodiment of a fleet operating system and its data flow. In an example embodiment, the fleet operating system and fleet intelligence system perform feedback on task run-time iterations of configuration activities, such as to adapt execution instances of a task execution plan. The embodiment of FIG. 136 illustrates an embodiment of the methods and systems of the robotic fleet platform 12002 shown and described herein, wherein feedback within the task configuration system 12018 includes task definitions 12304D and workflow definitions ( Facilitates repeating configuration activities when creating components of a task execution plan 12310, such as 12306D). As described for this embodiment, fleet intelligence layer 12004 may be used for at least this iteration. However, it is contemplated that the resources of the fleet intelligence layer 12004 may also or additionally be used to enhance execution of the task execution plan 12310.

In the example of FIG. 136 , job execution system 12022 of fleet operating system 12002 may receive a job execution plan 12310 from job configuration system 12018, for example, in response to a job request. Task execution system 12022 performs tasks by phasing out plans, activating and monitoring robotic units and other fleet resources, and providing feedback 12322, optionally real-time feedback, for example, based on robotic unit monitoring data. It may facilitate the execution of the execution plan 12310. This feedback 12322 may be processed by the artificial intelligence capabilities of the fleet intelligence layer 12004 to determine adjustments to the work execution plan, for example, task definitions, etc. When such feedback and adjustments are made in real time or near real time (e.g., in advance of upcoming task execution activities, such as steps in workflow 12306D), the functionality of task organization system 12018 is consistent with that of task execution system 12022. It can be iterated to modify an existing task execution plan, such as the instance of the plan currently being executed. In the building ventilation inspection example for the work execution plan iteration, the task of entering the ventilation system may involve removing ventilation portal covers from multiple locations within the building. Based on task run-time feedback from the robot (or team of robots) removing the initial ventilation portal cover from the ceiling port, the definition of this task is to keep the cover in place while removing the fasteners. can be adapted to require different maintenance techniques. In embodiments, the feedback may include images and/or videos of the removal task. In embodiments, the feedback may include a measurement of the weight of the cover as determined by the robot(s) performing the removal task.

This real-time (or near-real-time) visual feedback can be analyzed by the fleet intelligence system to determine, for example, that a portion of the baffle on the cover was deformed during removal. The artificial intelligence system in the fleet intelligence layer 12004 may perform a simulation of various cover support technologies and recommend one or more as input to the task composition system 12018 to update the corresponding task definition. In an embodiment, the fleet intelligence system sends an alert to the fleet operating system (12314) regarding the need to adapt this task definition, which can be used by the system to update pre-configured task definitions stored, for example, in the robot task library 12314, etc. 12002). Such alerts may be used by the fleet operating system to coordinate with the job execution system 12022 to ensure that pending ceiling-based vent cover removal tasks are not executed before being refreshed in the job execution plan 12310. In an embodiment, job organization system 12018 may release only a portion of job execution plan 12310 to job execution system 12022, such that the unreleased portion is adapted; This mitigates impacts to the task execution system, such as requiring the task to be halted, delayed, or otherwise corrupted while updates to the execution plan are made.

Examples of task composition, etc., presented in this application generally consider a single task being configured by task composition system 12018, but there may be many tasks being configured simultaneously. The methods and systems for real-time or near-real-time feedback described herein can be applied to any instance of work configuration activity being performed, such that feedback on the task definition of a first job will be beneficial to the task definition of a second job. while maintaining the necessary task isolation requirements to support concurrent processing of task requests from different entities (e.g., task identification data can be obfuscated).

136 also illustrates, among other things, a configuration for handling future work requests by selectively capturing data indicating completion of requested work as a form of feedback for use by fleet intelligence layer 12004 for learning and optimization. Means for further enhancing activities (e.g., task and fleet composition as described herein) are shown. In an embodiment, capturing data indicating completion of a requested task may include extracting such data from task completion data set 12326. This task completion data set 12326 may be structured to facilitate identifying information that may be useful for learning and optimization 12324. In one example, a work-complete data set may be designated, such as by the use of metadata tags, logical and/or physical separation, or other indicative data that indicates an exception or significant deviation from expectations. In an example, upon completion of a task, the count of repetitions of a robot function (eg, articulated arm movements to remove debris from a building ventilation system) may exceed the expected number. These excessive iteration counts may be flagged as candidate information for learning and optimization feedback 12324 to be extracted and sent to the fleet intelligence layer 12004. In an embodiment, the task execution plan 12310 may consist of an indicator of the type of data to be collected and used for learning and optimization feedback 12324. Fleet intelligence layer 12004 may recommend to work configuration system 12018 the types of data to be so presented based on other factors known to the fleet intelligence system, such as inquiries made by the robot design engineering team, etc. In embodiments, learning and optimization feedback 12324 may be used by the fleet intelligence layer to perform optimization of artificial intelligence services (e.g., recommending robot teams, robot types, workflows, etc.), among other things. there is. Referring to the description herein, pre-configured tasks, robot configurations, team configurations, etc. may be retrieved from library 12314. When this pre-configured aspect of the task execution plan is executed, data representative of performance may be flagged for use as learning and optimization feedback 12324 to continuously improve this pre-configured aspect. The results of using this data include pre-configured tasks b adapted to field conditions that can perform better in the real world. Other results of using this data include improved digital twins and machine learning models.

137, embodiments of task parsing system 12302 and task definition system 12304 are shown in interconnected block and data flow diagrams. The job description to be parsed may include relevant job description details, goals, objectives, requirements, preferences, etc., and may be described elsewhere in this application. Not all relevant task information may be included within the request, but one or more links to auxiliary task description data 12404 may be included. Auxiliary task data 12404 may be stored remotely from the task request data set (e.g., accessed via an Internet URL in the task description). Optionally, auxiliary operational data 12404 may be stored in a data structure accessible to fleet management platform 12000, such as a fleet library 12314, a requester-specific repository, etc. Auxiliary task data 12404 may include official standards (e.g., local disturbance regulations, safety (OSHA), electrical (NEC), quality, etc.), acceptance requirements (e.g., form, steps, timing, dependencies on other tasks). etc.), legal requirements of the work (e.g. union approvals, relevant laws, etc.), details of the work, standards for the requester's work (e.g. proficiency standards for the requester), industry norms (e.g. working hours, material selection, templates, etc.), approved vendors (e.g., from whom to obtain supplies and other consumables), references to pre-configured tasks, user interface templates/menus/screens for each aspect of the task ( For example, a way for users to request status, observe activity, change task requirements, respond to queries, etc.). Work request data and, if indicated, auxiliary data 12404 are processed by a task definition collection facility 12402 that operates in conjunction with a task data transformation module 12403 to generate task instance-specific content 12408. . This task instance-specific content can be, inter alia, defined in the input data (e.g., "perform task A before task B") and/or derived from it (e.g., the installation of an object must May include initial sequence timing (must be done after the object is received). Job data conversion module 12403 may interact with data processing system 12030 in converting job description data to utilize information derived from a fleet management platform accessible library, such as job and fleet library 12314. Collection facility 12402 may store some task description content, such as task identification information, links to internal auxiliary data, etc., directly into task instance repository 12408.

In embodiments, one or more human interaction capabilities to facilitate task parsing and task definition include interacting with a human (e.g., via text input, chat-bot, haptic-input, etc.) to retrieve task and task data. may include a knowledge-based system (e.g., AI-based, etc.) that can collect information for pre-formatting, organizing, and researching. This interaction can replace or supplement receiving job descriptions. By way of example, a job description may include reference to performing the task after normal working hours, which may include working after sunset. Disambiguation, such as the ability to examine interactive task descriptions, as mentioned here, can be beneficial in parsing task descriptions, and task definitions can include, for example, whether the task will require human-friendly lighting and, if so, what conditions this will require. You can decide. Because robot sensing may not require such lighting (for example, robot vision functions may be met through the use of infrared or other non-human visible light emissions), human-visible lighting may be required at certain times during task execution. It may only be required to be deployed at the beginning of a function (e.g., at the beginning of a function, when a delivery is taking place, when a human inspector is on-site, etc.). By providing the ability for human interaction as part of task parsing, such questions can be asked and answered interactively.

Job data transformation module 12403 may use job description information generated by or passed through collection facility 12402 to construct job instance content appropriate for the task definition. Operational data transformation module 12403 may use information provided by collection facility 12402 to query content within library 12314 (e.g., optionally via data processing facility 12030 as shown). there is. Content within the library that may be useful or informative for task definition includes task syntax (e.g., “front end loader”, “cybersecurity”, “hi-lift jack”). terms related to a given task, task type, set of tasks, etc.), robot type, robot capabilities (e.g. by type, cost, availability, etc.), keyword-to-task cross-references, workflow definition rules, May include task execution plan format/content/structure. Additionally, the library provides example multipurpose robot configurations (e.g., based on task keywords, etc.), example team configurations (e.g., to perform tasks of a particular type or class), task definitions, workflows, and workflow definitions. , example task execution plan(s), etc., may include templates for various task definition-related activities.

Keyword-based task lookup module 124010 can retrieve information, such as task-oriented keywords, from job instance repository 12408 and apply them to libraries 12314 to potentially identify pre-configured or templated tasks or portions thereof. there is. As an example, a job description may include keywords such as “submersed,” which may suggest a need for a robot that can perform tasks when submerged. When these keywords are combined with the action “submerged excavation,” the keyword-based task lookup facility 12410 can identify robot types that can perform excavation and be submerged. If a task's descriptor in the library aligns with one or more task description keywords, the task may be considered a candidate task for the task.

In an embodiment, task definer module 12412 processes candidate tasks provided by task lookup module 12410 as well as information in job instance repository 12408 to define tasks to be performed by one or more robots ( 12304D) can be formed. Defining a task may include tasks that are predefined by standards, laws, etc. As an example, a candidate task may include opening a manhole cover on a public sidewalk. Predefined tasks to meet standards and/or laws associated with such candidate tasks include notifying local code enforcement ordinances, local public utilities, placing safety signs at designated distances from open holes, marking open holes, etc. This may include keeping a watch on the hole while it is open and actively preventing unauthorized human entry. Each task definition can contain information useful for identifying the type of robot to perform the task.

In an embodiment, task definition system 12304 derives work requests (e.g., as provided by work request parser 12302) in the context of the robot type by identifying characteristics of the robot type that align with the task data. The task data can be processed. In an example embodiment, task definition system 12304 may determine that the task data indicates that the characteristics of the robot for performing the task may include nuclear radiation tolerance (e.g., the task of inspecting a nuclear reactor core). You can. In this example, task definition system 12304 may create a task definition 12304D for the reactor core inspection task that includes at least requirements for robot selection based on these characteristics. In this example embodiment, task definition 12304D may further include a required degree of tolerance for nuclear radiation (number of rads, duration of exposure, etc.). Task definition system 12304 may further determine characteristics of one or more robots that may not be suitable for inclusion in a single robot/robot type (e.g., based on task information derived from a work request). This determination may be based, for example, on robot characteristics and type data accessible in library 12314. In this example, task definition system 12304 may define multiple tasks, each having robot characteristics that match robot characteristic information in library 12314. In an embodiment, task definition system 12304 may optionally configure one or more tasks that require each type of a number of incompatible robot characteristics that fleet configuration system 12020 may use when configuring fleet resources, such as robots, etc. In addition to the representation of parts, tasks can be defined with a number of potentially incompatible robot characteristics. In an embodiment, task definition 12304D may be based on, for example, an ordering of task requirements (e.g., derived from task information in a work request), robot characteristics, and robot types that may be available in library 12314. , may include one or more suggestions for the type of robot to perform the task. As described below, fleet configuration system 12020 may evaluate task definition 12304D containing any proposed robot type. Other example data that may be communicated when defining a task may include task sequence dependencies that may be suitable for defining a workflow that includes the defined task. By way of example, a sample preparation task may be required to be performed after a sample collection task. These dependencies are documented in the sample preparation task, and the workflow definition system 12306 can rely on them. Task definer module 12412 may store defined tasks in a task instance repository, where defined tasks can be cross-referenced to task description data (e.g., keywords, etc.), such that future Detection can quickly result in appropriate task definition.

Figure 138 illustrates an example embodiment of a fleet configuration system 12020 in accordance with some embodiments of the present disclosure. In an embodiment, fleet configuration system 12020 provides specific software, hardware, and multipurpose robot configuration requirements for completion of a task execution plan. An example configuration of a fleet configuration system 12020 to provide these requirements is shown in the block diagram of FIG. 138. In this example configuration, fleet configuration proxy module 12466 may be configured to receive task definition 12304D from task configuration system 12018. The fleet configuration proxy module 12466 may be instantiated in connection with the processing of work requests by the work configuration system 12018 to facilitate access to and use of fleet configuration system 12020 resources and systems. These and other instantiations of fleet configuration proxy modules are further described in this application with respect to task configuration system 12018. Fleet configuration proxy module 12466 may process task definitions and forward them to fleet resource identification systems, such as fleet robotic operational unit identification system 12454 and fleet non-robotic operational unit identification system 12452. Each of these identification systems can process task definition data provided through a fleet configuration proxy to separate operational data from fleet resource data. The task definition defines the fleet resources required to perform the task, such as the type of robotic operating unit (e.g., one or more special-purpose robots) and support resources (e.g., power systems, lighting, communication systems, etc.). A set can be described. Robotic operational unit type identification system 12454 may provide task-specific robotic operational unit demand data 12476 to fleet configuration scheduler 12468. Task-specific robot operating unit demand data 12476 may identify the type and quantity of robots, specific robot operating unit capabilities (e.g., by unique identifier), robot operating unit capabilities, etc.

In some embodiments, fleet configuration scheduler 12468 may respond to work requests by allocating flit resources to meet work request demands. These requests may be preprocessed by task organization system 12018 and specifically by task definition system 12304 to facilitate fleet organization, allocation, and scheduling, as described herein. The fleet configuration scheduler 12468 processes inputs describing fleet inventory, such as robotic operating configuration unit inventory 12460 and non-robotic operating configuration unit inventory 12458, to identify candidate inventory elements for satisfying work requests. . This inventory may be adjusted based on existing allocations of robotic and non-robotic operating units. As an example, all special purpose robots of the type identified in the robot operation organization unit task specific demand data 12476 may be assigned throughout the duration of time during which the requested task is constrained to be performed. The Fleet Configuration Scheduler 12468 (e.g., with assistance from other platform resources such as the Fleet Intelligence Layer 12004, the Fleet Provisioning System 12014, etc.) Based on the available robot type equivalence data, a multipurpose robot can be assigned to an activity requested to be performed by a special purpose robot. To accomplish this assignment, fleet intelligence layer 12004 may include information describing the functionality to be provided by the special purpose robot indicated in task-specific demand data 12476 and the tasks and/or tasks required to be performed by the special purpose robot. Information describing the activity may be provided. Other context may also be available to the fleet intelligence layer 12004, such as differences in specifications for performing a task by a properly configured multipurpose robot and by a special purpose robot. Through the use of artificial intelligence, which may include determining the impact on the overall work request based on the use of two different robot types, fleet intelligence 12004 may provide robot replacement guidance to fleet configuration scheduler 12468. You can. This guidance may result in the allocation of multi-purpose robots and necessary configuration data/features (e.g., end effectors, etc.) for use in executing work execution plans corresponding to work requests that prompted these fleet configuration scheduling activities. In an example of fleet configuration scheduling, a 3D printable robot or a fleet service resource (e.g., a 3D printing factory or a third-party provider) may select robot parts that enable a general-purpose robot to perform the functions of a special-purpose robot (e.g. , can be assigned to the task of printing 3D printed robotic arms/end effectors as flexible/soft structures that can conform to irregular geometries to perform tasks.

In embodiments, task definition 12304D may include recommendations for one or more types of robots (e.g., based on task requirements, robot characteristics, and sorting of robot types), with the preferred type being the task definition. Can be specified in (12304D). By way of example, a task may be suitable for performance by a multipurpose robot or special purpose robot (e.g., robot characteristics aligned with task information may be found in library 12314 for configuration specific multipurpose robots and special purpose robots). there is). Although a general-purpose robot may be suitable, special-purpose robots may be preferred due to other factors in the task request, such as cost, availability, etc., and accumulated error thresholds that may be exceeded by the use of a general-purpose robot. When a multipurpose robot type is indicated in task definition 12304D, a reference to configuration data (and/or the data itself) may also be communicated in task definition 12304D.

As mentioned above, task information can be converted into task definitions that may require different or at least multiple robots. As an example, a sampling task requiring a robot with different characteristics to be defined may be identified as SAMPLE-T1. A first robot may be assigned by the fleet configuration system 12020 for a first portion of the task (e.g., Sample-T1-A for sample site preparation activities, such as removing objects interfering with sample motion). and a second robot/robot type may be assigned for the second portion of the task (e.g., Sample-T1-B for sampling activity), etc. When at least two robotic units are identified in a task, a task team designator may be communicated. By linking a team designator to a task identifier, fleet configuration system 12020 can take into account the specific needs of team members to perform a task when preparing fleet resource allocations for task execution.

Fleet configuration scheduler 12468 may rely on other fleet systems, such as fleet provisioning module 12014, which may contribute to and/or determine provisioning of fleet and third-party resources and supplies.

Other fleet systems, including the platform 12000 intelligence layer 12004, the fleet provisioning module 12014, and the fleet configuration scheduler 12468, are configured by the fleet configuration scheduler 12468 when configuring a fleet in response to task configuration activities, etc. May interact with a fleet configuration modeling system 12474 that may facilitate creation of fleet configuration options 12472 that may be considered. Fleet configuration modeling 12474 may provide a simulation of the fleet configuration, such as by using a fleet digital twin, which may optionally be associated with a digital twin system of the fleet intelligence layer 12004.

In embodiments, fleet composition scheduler 12468 may rely on fleet team organizer module 12470 to assist in determining/affecting team composition. Job-specific demand data 12476 may identify (e.g., recommend) a set(s) of robot operating units to be organized as a team. Additionally, task-specific demand data 12476 may represent information that may indicate team composition, such as the juxtaposition of robots performing a task. Team organizer 12470 may identify and/or specify team metadata for use when organizing a fleet. Team metadata may indicate team membership and a time frame for membership (e.g., by date, from the start of a task until the task is completed, etc.).

The fleet configuration scheduler 12468 generates the fleet robot operating configuration unit allocation data set 12462 and the fleet non-robotic operating configuration unit assignment information based on the configuration(s) generated for the provided task specific demand data 12476. A fleet allocation data set (which may be used by the fleet resource allocation and/or reservation capabilities described herein), such as data set 12456, may be updated. Various inputs, including fleet configuration affecting external data 12464 (e.g., weather, location data, traffic data, industry standards, task-specific contextual information, etc.), can be accessed through, among other things, the fleet configuration proxy 12466 may be optionally processed iteratively by the fleet configuration scheduler 12468 to generate a fleet configuration 12478 that may be returned to a running instance of the task configuration system 12018.

Figure 139 illustrates an example embodiment of a workflow definition system 12306 according to some embodiments of the present disclosure. In embodiments, workflow definition system 12306 may be configured to utilize resources of a fleet management platform to create definitions of workflows for requested tasks. The configuration of workflow definition system 12306 includes task definitions 12304D, which may be provided from task definition system 12304 or sourced from library 12314, and fleet configurations (e.g., via fleet configuration proxy 12305). It may include a collection module 12502 that receives and processes job specific fleet configuration information 12504 that may be provided from job configuration system 12018 interaction with configuration system 12020.

Gathering task definitions and/or fleet configuration information may include aligning fleet configuration information 12504 with one or more task definitions 12304D. As an example of aligning tasks with flit configuration information, fleet configuration information may be tagged as applying to one or more tasks within a set of collected task definitions, such as an identifier for a task or tasks. Another way to align task definition(s) with fleet configuration information may be based on the timing of such collection, such that, for example, when a fleet configuration reference/value is received concurrently with a task definition, the collection module 12502 may Two data items can be marked as sorted. Other ways to align task definition(s) with fleet configuration information may include one or more data values in the task definition, which may be a data set, a linked list, or a data set indicating the fleet configuration information by which the task(s) should be aligned. It can be a flat file, structured data set, etc. The fleet configuration information may include one or more task identifiers to which the fleet configuration information is associated and/or should be applied when creating a workflow definition.

Collection may further include processing references (e.g., URLs, hyperlinks, external names, etc.) to workflow content within library 12314 that may be found in any of the collected content. In one example, the task definition may include the name of the task stored in library 12314. The collection module 12502 may provide syntax (e.g., a prefix may be added to a task identifier indicating that a task will be retrieved from a library) and/or task definition structuring (e.g., a subset of tasks may be retrieved from a library). The name can be identified by a list of task names stored within a subset of the task definition structured to indicate that the task name is Examples of collections in this application relate to instances of collections of one or more task definitions, but collections may be performed on batches of tasks. Multiple instances of collection module 12502 may be instantiated and concurrent operations may be performed to process multiple task definitions. Optionally, a stream of task definitions can be received by collection and each task in the stream is collected sequentially.

One or more results of processing by collection module 12502 may determine dependencies between tasks, e.g., task dependencies, which may determine which tasks need to be performed sequentially and which tasks can be performed independently of other tasks. A set of workflow definition activities including decision module 12506 may be presented. Task dependency determination module 12506 may also determine the dependency of a task on other factors such as availability of fleet resources, calendar/date/time, readiness of supply materials, etc. Dependencies on other arguments can be identified in the task definition, for example, by marking a given task state as a starting point for the task. In an example of a task state task dependency, the task of processing a sample of material may depend on the material being received by a sample catalog robot, etc. Additional other factors task dependencies may result from task definitions given during collection (e.g., based on aligning tasks with fleet configurations that establish dependencies on the availability of fleet resources such as special purpose robots, etc.).

Task grouping activity 12508 processes the results of task dependency activity 12506 such that tasks that depend on a given task being completed (e.g., opening a building ventilation system port) will be grouped for concurrent execution. As you can, you can create task groups based on scope criteria. Task grouping may be based on dependency on fleet resource availability, such that tasks dependent on fleet resources can be grouped and performed once the resources become available. The execution order of these grouped tasks may be based on dependencies between tasks. Typically, tasks are used for various purposes such as cost reduction, resource protection, task prioritization, available task execution funds, expected fleet resource maintenance needs, earliest task start/finish time, latest task start/finish time, etc. Can be grouped.

Task workflow step definition activity 12510 may determine which task(s) can be organized into each step of one or more workflows. Based on inter-task dependencies (or lack thereof), multiple workflows may be defined, each workflow including one or more workflow steps defined in workflow step definition activity 12510. As an example of inter-task dependencies, prohibited tasks, such as those required by electrical safety standards, can serve as reference points to which other workflow development activities must adhere. Referring back to the building ventilation system inspection example referenced in this application, the set of workflow steps for opening ventilation ports may consist of multiple workflows, one for each ventilation port (selection based on other conditions). accompanied by adaptive adaptations). Additionally, a workflow step, once defined, can be assigned and/or referenced to multiple workflows. When a dependency exists, such as the availability of a special purpose robot to perform a task in a workflow step, the multiple workflows themselves may be made dependent. In the example, when the task of opening a ventilation port is defined for a special purpose robot and the task requires opening four ports, the workflow containing this port opening task will open each task only when the required resources are available. This can be done dependently for the flow to start. Performance of other tasks in this workflow may be concurrent even if the initial task of opening the port must be performed sequentially due to fleet resource utilization dependencies.

In embodiments, the defined workflow steps may be adapted variations of candidate workflow steps 12514, such as workflow steps retrieved from library 12314. Workflow step definition activity 12510 may be performed by a data processing system 12030 and/or an artificial intelligence service, such as 12028, to adapt candidate workflow steps for use in defining one or more workflow steps for a given task. Input may be requested from other fleet resource platform services.

Information such as workflow step dependencies may be utilized by workflow step integration activities 12512, which may receive step integration recommendation(s) 12516 from fleet intelligence layer 12004, etc. Workflow step linkage activity 12512 may create a data structure representing a sequence for performing a defined workflow step (e.g., workflow definition 12306D). A workflow definition (12306D) captures task-specific workflow information such as workflow step sequencing, workflow step performance sequence, workflow step independence, step-by-step links to workflow steps, workflow success criteria, cross-workflow dependencies, etc. Can contain data.

In embodiments, workflow definition(s) 12306D may be stored in a job instance repository 12408 where they may be referenced as needed during job configuration and/or job execution. These may be stored in a fleet library 12314 where they can be referenced by other tasks, third parties such as task requesters, etc. They may be stored elsewhere (e.g., a cloud storage facility) based on architectural considerations, such as distributed on edge computing infrastructure resources in close proximity to the work deployment site, etc.

In embodiments, workflows may be simulated as indicated in the description of task configuration system 12018. The results of the simulation may relate to the collection module 12502, where collection operations may be improved, for example, sorting of fleet configuration data and task description data. Results may also be communicated as feedback 12406 to other components of platform 12000 to improve task definition, job configuration, fleet configuration, etc.

In a specific example, an exemplary robotic fleet task may include inspecting a building ventilation system. Work request parsing system 12302 may parse the work request and any related documents to identify ventilation system inspection routines, tasks, actions, steps, requirements, etc. The task request parsing system 12302 may provide the parsed information to the task definition system 12304. In embodiments, the inspection procedure associated with the work request may indicate one such inspection procedure step for entering the ventilation system (eg, through a wall or ceiling register, etc.). Task definition system 12304 may identify a plurality of tasks associated with the procedural steps of entering the ventilation system. These tasks may include: collecting information about the physical configuration of the ventilation system to identify the location and type of registers available in the building; analyzing the physical information of the ventilation to select candidate registers; requirements for accessing the register (e.g., whether it is located behind a locked door, whether the robot needs to be lifted to enter the system through the register, etc.), and requirements for removing the register's cover/grate. Deciding on tools, etc. Additional information that may be relevant to one or more of the tasks for this procedure step include, without limitation, size limitations of the robot entering the ventilation system (which may not be specified in the procedure, but may be based on ventilation system entry ports), may require decisions as tasks, such as those based on information about the physical configuration, weight limitations of one or more robots, etc. In embodiments, tasks defined by task definition system 12304 may be performed by fleet resources, including human fleet resources, resources other than individual robotic operating units, such as digital twins, etc., that may operate on a platform processing system. May include data analysis tasks. Other routines/tasks for entering a ventilation system that may require definition may include orienting the robot for entry. The result of these decisions may be to add requirements for the robot to perform the task(s). In embodiments, vertical entry may require a ventilation duct gripper oriented at the front of the robot. Task definition 12304D may include details, such as duct gripper orientation, that other systems on platform 12000, such as fleet configuration system 12020, can use when configuring aspects of a fleet. Typically, a discrete robot task definition 12304D is a plurality of (basic/elementary/generic) robot movements and/or routines that are optionally sequenced and aggregated together to satisfy the low-level goals (e.g., tasks) of the robot fleet operations. May include (explicitly or implicitly). Accordingly, task definition system 12304, which creates task definitions 12304D for specific robot fleet tasks (e.g., inspecting a ventilation system as illustrated herein), may provide, for example, access to a ventilation system. By aggregating and/or adapting such robot movements to satisfy some criteria for performing a target task, such as removing a panel, a task definition can be created that implements more than typical robot element movements. Robotic actions such as positioning and rotating fasteners, grasping access panels, positioning removed panels, reserving fasteners, etc. are common robot routines or movements that can be aggregated and adapted into task-specific tasks. You can take it in. These general robot routines or movements may be available to task definition system 12304 to facilitate defining relevant aspects of a task based on task requests and associated criteria. In the example of inspecting a ventilation system, positioning a fastener on an access panel may be adapted during the operation of this task or adapted during the task based on details of the target access panel that can be identified in the task definition 12304D. For this purpose, it can be entrusted to an intelligent system such as a robot-based intelligent system. Basic robot actions, such as rotating and removing a fastener, can be adapted based on information provided in the task definition, which can define appropriate end effectors, torques, and travel distances. In embodiments, this adaptation may be left to the robot control function to determine simultaneously with performance of the task which end effector, etc. to use. Information within task definition 12304D may facilitate coordination of the robot to grasp the access panel. This information may include the orientation of the panel, the weight of the panel, the characteristics of the panel, the size of the panel, etc. to ensure safe gripping of the panel while avoiding damaging the panel. The task/action of placing the removed panel ensures that objects within the task location (e.g. furniture, windows, walls, etc.) are not damaged by the panel and that the path through the task location is not blocked or hazardous to humans. It can be configured with a certain degree of location-specific flexibility to reserve the robot motion control system for other criteria (e.g., safety standards and practices, workplace policies, governance, etc.) that can be utilized. Accordingly, these tasks may be interpreted by the fleet configuration system 12020 such that robots can be matched to defined tasks, including features for evaluating deployment locations, such as vision systems. In embodiments, this flexibility may be selected from robot configuration library 12314.

Continuing further to describe the exemplary robotic fleet task of inspecting a building ventilation system, workflow definition system 12306 may be used to establish workflows for at least the procedural steps of entering the ventilation system (e.g., optionally Information output by the task definition system 12304 and the fleet configuration system 12020 (via the fleet configuration proxy 12305) may be collected. At a level of abstraction, this procedural step may involve two main tasks: (i) removing the access panel, and (ii) entering the ventilation system. Information from task definition system 12304 may indicate that task (i) is a prerequisite for performing task (ii). Accordingly, workflow system 12306 may define a workflow in which task (i) occurs before task (ii) for this portion of the requested work. Additional task (iii) may include capturing 3D images of the environment in which entry to the ventilation system is taking place. Information from fleet configuration system 12020 regarding one or more robots configured for this task indicates that two robots are configured, a first robot for task (i) and a second robot for task (ii). can do. The workflow system uses this information to define a workflow that causes the second robot to perform task (iii) while waiting for the first robot to complete task (i), thereby creating tasks (i), (ii), and It may be determined that the order of (iii) can be optimized. If the fleet configuration information for these tasks indicates that a single robot will be provisioned for these three tasks, then the workflow system will order the tasks to be (iii) followed by (i), then (ii) It can be defined as: These alternative workflow configurations, responsive to information provided to the workflow system, indicate the degree of flexibility of the workflow system in defining workflows, for example, to ensure efficient use of fleet resources.

Simulation of the workflow of these three tasks through workflow simulation system 12308 can also provide insight into any of the task definition, fleet resource allocation, and workflow definition. As a non-limiting example, a simulation of a workflow defining the sequence of tasks as (i), (ii) and finally (iii) can be performed in step (ii), with a single robot operating unit performing these three tasks. Since it will be placed inside the ventilation system, it can be calculated that step (iii) cannot be performed for a single robot assignment as indicated. The results of the simulation can at least be fed back to the workflow system to rework the workflow. In embodiments, data arising from the simulation (e.g., failure to perform step (iii)) may be fed back to any previous steps in the work configuration system process, such as task definition, fleet configuration, etc. In another example of a workflow simulation, if two robots are configured to perform this task as described above, the workflow would be such that the first robot removes the access panel (task (i)) at the same time that the second robot If 3D imaging of the task area (task (iii)) is required, the simulation may attempt to perform a simulation of the 3D imaging function with, for example, a digital twin of the second robot. The simulation may fail if a second robot with 3D imaging capability is not configured by the fleet configuration system. Feedback from such simulations can result in a range of changes in task configuration. Two example changes may include: (i) adjusting the robot configuration (changing the configuration of the second robot to maintain workflow and include 3D imaging capabilities); and (ii) coordinating one or more task assignments (assigning 3D imaging functions to a first robot and coordinating workflow).

In an embodiment, the task execution plan 12310 for inspecting a building ventilation system includes at least three defined tasks (i), (ii), and (iii), fleet resources (e.g., robots) for each task. configuration) and allocation information (e.g., from fleet configuration system 12020), and a workflow defining a sequence of three tasks.

In light of the foregoing disclosure, fleet management platform 12000 may be a standalone service or may be integrated into a larger system-of-systems (SoS). Additionally, fleet management platform 12000 is configured to facilitate many different types of fleets for different types of tasks. In addition to the configurations described above, some additional examples of fleet and robotic operational organizing units that may be configured by fleet management platform 12000 are provided below.

FIG. 139 illustrates an example embodiment of a multipurpose robot 12100 according to some embodiments of the present disclosure and may be applied to the general example of MPR 12100 of FIG. 129. Generally, multipurpose robot 12100 is designed, built, configured, and operated to maximize operational flexibility in individual and group deployment scenarios. In this way, multipurpose robot 12100 can be configured and reconfigured to perform specific task-specific functions in addition to the baseline functions of multipurpose robot 12100. In embodiments, MPR 12100 may be configured to operate autonomously, semi-autonomously, or using instructions provided by one or more users. In an embodiment, MPR 12100 may include a baseline system 12102, a module system 12120, a robot control system 12150, and a robot security system 12170. For task-specific capabilities, MPR 12100 may integrate configurable and interchangeable hardware and software modules provided by physical interface module 12122 and control interface module 12130 of modular system 12120. These modules may be mounted on and interface with the control system 12150, robot security system 12170, and/or baseline system 12102 as required for robot mobility, power distribution, etc.

In an embodiment, the baseline system 12102 of the MPR 12100 includes various hardware, devices, interfaces, processors, software, and systems that perform the baseline functions of the MPR 12100. In some embodiments, baseline system 12102 may include an energy storage and power distribution system 12104, an enclosure that surrounds some or all of the components of MPR 12100, which stores energy and delivers power to other components of the robot. 12106), an electromechanical and electrofluidic system 12108 that operates and controls the mechanical components of the MPR 12100, a transportation system 12110 that includes mechanical components that physically move the MPR 12100 in the intended environment, a base. a vision and sensing system 12112 that includes a baseline set of sensors used in connection with performing line functions and/or specific task specific functions, and one or more skeletal components configured to provide shape and structure to the MPR 12100 It may include a structural system 12114 including.

As can be appreciated, the baseline system 12102 of the MPR 12100 is designed to operate the MPR 12100 in a particular operating environment or condition, regardless of the tasks the MPR 12100 may be customized to perform. It can be configured according to the characteristics required (for example, for operation in heat, low temperature, humidity, land, sea, underwater, air, underground, etc.). Accordingly, different classes of MPRs 12100 configured for operation in different operating environments or conditions will have different configurations of each baseline system 12102 of the MPRs 12100. For example, an example baseline system 12102 of a four-leg ground MPR 12100 designed to operate on solid ground in wet weather conditions may be stored, for example, in a battery and supplied by a wireless power distribution system 12104. An IP-43 rated enclosure 12106 housing four separate mechanical legs having an electric motor 12112 on each leg 12110, driven by electrical energy. In another example, an exemplary baseline system 12102 of an underwater robot MPR 12100 designed to operate underwater is an IP 68 accelerator that accommodates a waterjet propulsion system using an electric motor 12112 driven by electrical energy stored in a battery. May include a graded enclosure 12106. In another example, the third baseline system 12102 of an MPR 12100 designed for operation in mud may include a tracking wheel 12110, where power is coupled to a hoseless hydraulic power transmission system 12104. Powered by a gasoline engine.

In embodiments, the energy storage and power distribution system 12104 of MPR 2B00 may include a hydraulic system, an electrical system, a nuclear system, a supercapacitor, a flywheel, a solar or photovoltaic cell, a fuel cell, a battery, a power cord, a dynamic or It may include one or more power source(s) configured to power various components of MPR 12100, such as piezoelectric battery charging devices, inductive charging or wireless power receivers, and other types of power systems. In embodiments, the choice of power source may depend on different factors such as the size and shape of the MPR 12100, the environment in which the MPR 12100 is operating, the tasks the MPR 12100 needs to perform, etc. In embodiments, selection of power source may be based on these factors and may support a wide range of use case scenarios for MPR 12100. For example, the MPR 12100 switches to wall power for fixed-position applications that can consume significant power, for example, to move heavy loads in construction or earthmoving applications, but also to mobile devices responsible for house cleaning. While operating as a robot, it can rely on a lithium-ion battery system. In embodiments, different components of MPR 12100 may be powered by the same power source, may be powered by multiple power sources, or may be each connected to a different power source.

In embodiments, the power components within the energy storage and power distribution system 12104 may include a plurality of lithium-ion smart batteries and a rechargeable battery or battery pack configured to provide charge to other components of the MPR 12100. You can. The use of smart batteries allows for modular battery systems, potential upgrades as new chemistries become available, and monitoring of power system health at the individual battery level. Using multiple batteries yields a system that is tolerant to the failure of any single battery element because such loss only reduces the maximum available power and energy storage. In an embodiment, MPR 12100 may be driven by a primary power source comprised by an AC electrical supply grid from a power grid and a secondary source comprised by a battery pack. In embodiments, system power is provided by a fixed source external to the MPR 12100 using one or more power repeater coils, and an integrated wireless power distribution system provides power to subsystems of the MPR 12100, such as sensor packages. Provides, monitors, and manages flows and supplies.

In an embodiment, the power source within energy storage and power distribution component 12104 includes a hydraulic system configured to use fluid power to drive MPR 12100. The various components of the MPR (12100) are supplied with hydraulic fluid that is stored in a reservoir and delivered through high-pressure supply lines using pumps at specified pressures and flow rates to one or more hydraulic elements, such as, for example, various hydraulic motors, hydraulic cylinders, and actuators. It can operate based on The hydraulic system may transmit hydraulic power through tubes, flexible hoses, or other links between components of MPR 12100 by pressurized hydraulic fluid. The specific design and components of the hydraulic system may vary, and any number or combination of valves, control systems, actuators, reservoirs, pumps, or any other items may be included as desired. Typical response times for this type of hydraulic system are very fast, on the order of a few milliseconds or less.

In embodiments, hydraulic systems are designed to utilize additive manufacturing methods and their associated design advantages to produce manifolds and reservoirs that minimize hoses and connections that can lead to leaks and system inefficiencies. The hydraulic system may include the ability of the MPR (2B00) to apply repairs, service equipment and handle emergency situations that the application must avoid. In embodiments, hydraulic systems are designed to utilize additive manufacturing methods and their associated design advantages to produce manifolds, reservoirs, and distribution systems containing valve actuation.

In embodiments, enclosure 12106 of MPR 12100 may include any housing or other physical component that includes at least a portion of MPR 12100. The structure of enclosure 12106 may vary and may depend on the operation that MPR 12100 may be designed to perform. In an embodiment, enclosure 12106 is a rectangular metal box with an interior space isolated from the environment by an exterior wall having a predetermined environmental resistance. The internal space is an energy storage and power distribution system (12104), an electromechanical and electrofluidic system (12108), a transportation system (12110), a vision and sensing system (12112), a robot control system (12150), and a robot security system (12170). It can accommodate various components of the MPR (12100), including etc.

In some embodiments, the enclosure 12106 of the MPR 12100 may be designed for robustness and the ability to tolerate external environments. For example, protection can be provided against water, humidity, dust, vibration, and temperature. One or more sealing mechanisms may be provided to protect against water ingress. In some cases, a water-repellent coating may be provided. Accordingly, MPR 12100 can tolerate external weather conditions such as rain, wind, sun, or snow.

In some embodiments, the enclosure 12106 of the MPR 12100 complies with IP-68, indicating optimal protection against dust and water. IEC Standard 60529, the IP Code or Intrusion Protection Code, sometimes referred to as the International Protection Code, classifies and rates the degree of protection provided by mechanical casings and electrical enclosures against ingress, dust, accidental contact, and water. The IP rating is indicated by two codes: “IP (1st code)” (2nd code). The first symbol indicates the degree of protection of electrical equipment and cabinets against solid foreign objects, which has 7 degrees from "0", meaning no protection against dust ingress, to "6", meaning no dust ingress inside. It is expressed as The second symbol represents the level of protection against water ingress, which is expressed in 9 levels from “0” meaning no protection against water ingress to “8” meaning optimal resistance. When the grade cannot be determined, “X” is displayed.

In some embodiments, the enclosure 12106 of the MPR 12100 is made of non-conductive and heat-dissipating smart materials. The materials can help protect sensitive electronic components, including components of the vision and sensing system 12112 and the robot control system 12150.

In some embodiments, the electromechanical and electrofluidic systems 12108 of the MPR 12100 may include a set of electrical and mechanical components configured to provide form and structure and enable operation of the MPR 12100. A set of electrical and mechanical components can work together to enable MPR 12100 to perform a variety of functions. For example, electrical components may be configured to provide power from a power source within the energy storage and power distribution system 12104 to various mechanical components. Electrical components may include a variety of mechanisms that can process, transfer, or provide electrical charges or electrical signals. Among possible examples, electrical components may include electrical wiring, circuitry, or wireless communication transmitters and receivers to enable operation of MPR 12100. Electrical components also include brushed DC motors, brushless DC motors, switched reluctance motors, universal motors, AC multiphase squirrel cage or wound-rotor induction motors, AC SCIM split-phase capacitor-starting motors, AC SCIM split-phase capacitor-starting motors. Working motor, AC SCIM split-phase auxiliary starting winding motor, AC induction shaded pole motor, winding-rotor synchronous motor, hysteresis motor, synchronous reluctance motor, pancake or axial rotor motor, stepper motor, or any other type. It may include an electric motor, including an electric motor, or a non-electric motor. Electric motors can help move one part relative to another. The mechanical component represents the hardware of the MPR 12100 that can enable robotic systems to perform physical movements. The specific mechanical components may vary based on the design of the MPR 12100, but may include some basic skeletal components, such as a structured body connected to one or more appendages or end effectors through one or more joints.

In some embodiments, MPR 12100 includes a structural system 12114 comprising a plurality of joints, appendages, and skeletal components configured to provide shape and structure to MPR 12100. Structural system 12114 may include a body, torso, head, legs, arms, wheels, end effectors, manipulators, gripping devices, etc. The skeletal component of structural system 12114 may include an inner core having male and/or female ends. The various skeletal components may be connected to enclosure 12106 and other skeletal components via joints, mechanical fasteners (e.g., nuts and/or bolts), actuators, hinges, latches, or other suitable mechanisms. Skeletal components of structural system 12114 may provide support and allow transfer of fluids, power, data, etc. Joints join skeletal components together and may allow movement in one or more degrees of freedom. Joints may allow skeletal components to move in vertical and horizontal directions as well as rotate relative to each other. For example, MPR 12100 may include one or more arm motors that may be used to move the arms relative to the body. In embodiments, the arm motor may be actuated by a source of energy, typically electric current, hydraulic fluid pressure, or pneumatic pressure, and may include an actuator that converts that energy into motion. Examples of actuators include linear actuators, solenoids, comb drives, digital micromirror devices, electric motors, electroactive polymers, hydraulic cylinders, piezoelectric actuators, pneumatic actuators, servomechanisms, servo motors, thermal bimorphs, screw jacks, or any other. Types of actuators may include hydraulic, pneumatic, electrical, mechanical, thermal, and magnetic types.

MPR 12100 may be comprised of zero or more legs or other movable or fixed bases depending on the particular application or intended use of MPR 12100. Implementations of MPR 12100 with zero legs may include wheels, treads, or some other form of locomotion. An implementation of a robotic system with two legs may be referred to as biped, and an implementation with four legs may be referred to as quadruped. Other implementations with 6 or 8 legs may also be possible. The structure of MPR 12100, including enclosure 12106, body, shape, size, skeletal components and materials, etc., may vary and may depend on the operation that MPR 12100 may be designed to perform. For example, when developed to carry heavy loads, MPR 12100 may have a wide body to allow placement of the loads. Similarly, when configured to reach high speeds, MPR 12100 may have a narrow, compact body made of lightweight materials.

In some embodiments, MPR 12100 may be structured to mimic the human body, such that MPR 12100 includes a torso, a head, two arms, and two legs. Actuators can act like muscles and joints and allow skeletal components to rotate relative to each other in a manner similar to bones in the human body rotating about joints. For example, joints may be configured to move skeletal components in a manner similar to movement of the hand, fingers, elbow, waist, knee, wrist, shoulder, etc. Construction materials may include biologically inspired artificial skin with sensors for detecting touch, acceleration, proximity and temperature.

In embodiments, the transport system 12110 of the MPR 12100 may include one or more body motors that can be used to move the MPR 12100 through one or more transport vehicles. A transport vehicle may be configured to facilitate movement of MPR 12100 across a surface. In some embodiments, the transport vehicle may include wheels, casters, treads or tracks, low friction pads or bumpers, low friction plates, skis, pontoons, or friction between the MPR 12100 and the surface on which it is desired to move. It may include any other suitable device configured to reduce . In further embodiments, the transport vehicle may include a propeller, a miniaturized jet engine, or any other air transportable device that allows MPR 12100 to fly or function similarly to a drone aircraft. In further embodiments, the transport vehicle includes fins, water jets, screws, or any other water transportable device that can enable MPR 12100 to be moved on or below the surface of water. can do. In further embodiments, the transport vehicle may include rockets, and ion drives, gyroscopes, or any other space transportable device that may allow MPR 12100 to travel in space.

In embodiments, vision and sensing system 12112 may include various sensors within MPR 12100 that act as input mechanisms to collect information from the environment. This sensing information is processed to operate other subsystems, including energy storage and power distribution system 12104, electromechanical and electrofluidic system 12108, transportation system 12110, and structural system 12114. Provided to robot control system 12150. Thereby, vision and sensing system 12112 enables MPR 12100 to monitor and navigate its environment, including interacting with and manipulating one or more objects within that environment. An example vision and sensing system 12112 is described in detail in conjunction with FIG. 142.

The robot control system 12150 includes various hardware, devices, interfaces, processors, software, and systems for controlling the operation and behavior of the MPR 12100. For example, control system 12150 may cause MPR 12100 to move to a specific location by following a path and avoiding obstacles in the path. As another example, control system 12150 may enable MPR 12100 to cooperate with others or interact with its environment, including grasping or manipulating one or more objects in the environment.

Robot control system 12150 may read from sensors to update actuators that act as output mechanisms for driving joints, arms, legs, end effectors, etc. Robot control system 12150 provides precise motion control of MPR 12100, including control of fine and gross movements necessary to manipulate objects. Control system 12150 is capable of controlling each robotic joint and other skeletal component of structural system 12114 independently by isolating it from other joints and skeletal components, but can also control multiple joints in performing relatively complex operational tasks. Multiple joints can be controlled interdependently to fully coordinate the action of

Robot control system 12150 may communicate with other systems in the MBR, other robots, and/or fleet management platform 100 via wired or wireless connections, and may be further configured to communicate with one or more users. For example, control system 12150 may receive input (e.g., from a user or from another robot) representing commands to navigate to a location. Accordingly, control system 12150 may serve as an interface between different components of MPR 12100, such as between sensors and actuators, between mechanical and electrical components, as well as between MPR 12100 and a user.

In an embodiment, the robot control system 12150 includes an intelligence layer 12140, a performance management system 12146, a task management system 12144, a data processing system 12142, a module management system 12148, and a communication system 12152. , may include and/or utilize a navigation system 12154, a safety and compliance system 12156, a motion planning system (MPS) 12158, and/or a controller 12160. It is understood that the foregoing description of robot control system 12150 is also applicable to other types of robots, including special purpose robots and/or exoskeleton robots.

In an embodiment, intelligence layer 12140 provides a framework for providing intelligence services and helps enable MPR 12100 to make decisions, predictions, classifications, etc. In embodiments, the intelligence layer 12140 may request requests from the robot control system 12150, the baseline system 12102 of the MPR 12100, etc. to provide certain intelligence (e.g., decisions, classifications, predictions, etc.). receives. For example, the intelligence layer may be responsible for making decisions about controlling the motion of MPR 12100 based on environmental data (e.g., maps, coordinates of known obstacles, images, etc.). In embodiments, the framework provided by intelligence layer 12140 may be configured as part of a broader intelligence layer extending to the fleet 4D00 and/or platform level, as described elsewhere in this disclosure. there is.

In an embodiment, intelligence layer 12140 may include an intelligence layer controller 12141 and an artificial intelligence (AI) service 12143. In embodiments, intelligence layer controller 12141 may be configured to determine the type of service to be provided by artificial intelligence service 12143 and, in response, determine the governance standards and/or standards to be applied by artificial intelligence service 12143. The set of analyzes can be determined. The intelligence layer 12140 of the MPR 12100 (or SPR or exoskeleton) may include some or all of the intelligence services 12143 of the intelligent system described above. Additionally, in some embodiments, robot-level intelligence layer 12140 may send intelligence requests to a higher level (e.g., fleet level, edge device, or fleet management) when MPR 12100 is unable to perform a task autonomously. It may be configured to transmit to the platform 12000). An example embodiment of robot-level intelligence layer 12140 along with its components and subsystems is described in detail with FIG. 140 .

In embodiments, performance management system 12146 is configured to manage the performance of one or more robotic resources including health, energy, heat flow, networks, etc. In embodiments, performance management system 12146 may provide thermal management services 12161, energy management services 12162, monitoring and notification services 12163, network management services 12164, and/or predictive maintenance services 12165. It can be included.

In an embodiment, thermal management service 12161 may use robot sensors, task historical data, ambient conditions, material properties, form factors, etc., and a set of acceptable results to drive optimization algorithms that manage heat flow in multipurpose robot 12100. can be used. This can be used to actively manage thermal conditions or optimize heat transfer to maintain acceptable operating conditions. In embodiments, thermal management service 12161 may help recover waste heat energy. For example, waste heat could be transferred to actively cool hotter components, such as for use with emerging nanoscale or other thermoelectric devices. In an embodiment, thermal management service 12161 may provide thermal information such as fins, vanes, biomimetic elements, mesh, fabric, fans, etc. in MPR 12100 in addition to robot sensor data, task historical data, ambient conditions, material properties, form factors, etc. The set of acceptable results may be utilized to drive optimization algorithms (e.g., quantum optimization algorithms and/or neural network optimization algorithms) that design and manage the operation of the transfer component.

In embodiments, energy management service 12162 helps intelligently manage available energy resources and maintain system capability while the robot operates in a dynamic operating environment. For example, if grid energy may not be available and the robot discovers that there is a need to conserve available batteries, MPR 12100's energy management service can replace one or more Energy storage and recovery devices can be activated. The device enables the MPR 12100 to harvest energy during the braking phase of the motor - this energy would normally be wasted -, store it, and provide it back to the system when needed. In embodiments, an energy sharing device may share the braking energy of a motor to drive another (non-braking) motor or actuator on a common network. In embodiments, energy management service 12162 may include machine learning-based predictive energy management that automatically activates energy harvesting and sharing devices and disables non-essential functions based on need.

In embodiments, monitoring and notification service 12163 may be configured to monitor and report on one or more conditions in MBR 12100. In some of these embodiments, monitoring and notification service 12163 performs summary calculations on tracking metrics of various resources to discover out-of-routine characteristics. In some example embodiments, monitoring and notification service 12163 may perform vibration analysis indicative of robot health, including the condition of one or more motors or mechanical components. In some of these embodiments, monitoring and notification service 12163 is a machine learning model that is trained to diagnose specific conditions in the robot (e.g., failed components, loose components, etc.) to predict the presence or likelihood of the specific conditions occurring. You can use . In embodiments, monitoring and notification service 12163 may utilize one or more machine learning models, including vision models, to monitor, discover, and predict emerging robot failure modes. In embodiments, monitoring and notification service 12163 may also provide alerts and notifications to users upon discovering any unusual characteristics. For example, when predicting that the battery is about to be completely depleted, monitoring and notification service 12163 may provide alerts and notifications to the user using voice messages. Additionally or alternatively, monitoring and notification service 12163 may be used to send notifications to a user's computing device (e.g., computer, tablet computer, smartphone, phone, mobile phone, PDA, TV, gaming console, etc.). Email, text messages, instant messages, phone calls, and/or other communications (e.g., using the Internet or other data or messaging networks) may be used. In embodiments, an error notification may provide the user with an option to stop the operation or make an adjustment to one or more settings associated with the error notification. In embodiments, monitoring and notification service 12163 may provide customized reporting to users, including analysis based on real-time and historical data regarding the status and/or diagnostics of various resources of MPR 12100.

In an embodiment, network management services 12164 include a set of assigned policies, procedures, workflows, and responsibilities to improve or maintain optimal network performance. In embodiments, network management service 12164 may evaluate network flow data, packet data, and network infrastructure metrics to identify and mitigate instances of bottlenecks or network issues that may impact the operation of MPR 12100. there is.

In embodiments, predictive maintenance service 12165 may cause one or more components or subsystems of MPR 12100 to perform maintenance based on simulated data derived from a digital twin system or real-world data derived from monitoring and notification 12163. You can predict when you should receive it. In embodiments, predictive maintenance service 12165 may access the intelligence layer 12140 of MPR 12100 to predict expected wear and failure of components of MPR 12100 by reviewing historical and current operational data; Thereby reducing the risk of unplanned downtime and the need for scheduled maintenance. For example, in an embodiment, predictive maintenance service 12165 may provide an intelligence request containing current operational data (e.g., sensor data, environmental data, etc.) obtained from MPR 12100 to the intelligence layer. and, whereby the intelligence layer 12140 (e.g., machine learning service) may use one or more machine learning models (e.g., prediction models, classification models, neural networks, etc.) to identify potential failures of components of the MPR 12100. ) can be used. In embodiments, a machine learning model may be trained using data regarding robot specifications, parameters, maintenance results, environmental data, sensor data, execution information, and notes to predict failure and perform predictive maintenance. Additionally or alternatively, the machine learning service may include clustering algorithms to identify failure patterns hidden in failure data to train models to detect non-characteristic or abnormal behavior. Failure data and historical records across multiple robots can be clustered to understand how different patterns correlate with specific wear behavior and develop maintenance plans that resonate with the failures.

In another example, the predictive maintenance service 12165 utilizes the digital twin services of the intelligence layer 12140 to form a digital twin of the MPR 12100 (e.g., in an environment in which the MPR 12100 is operating or will operate). The operation may be simulated, whereby the digital twin simulation may reveal potential wear of MPR 2B00 and/or potential failure of components of MPR 12100. In this example, over-servicing or over-maintaining the MPR 12100 can be alleviated, thereby reducing the cost of the MPR 12100 or its components by addressing these issues proactively or in a just-in-time manner. This can reduce costly downtime, repairs or replacements.

In an embodiment, task management system 12144 coordinates between intelligence layer 12140 and task execution systems of fleet operating system 12002, library 12314, vision and sensing system 12112 to execute tasks. Task management system 12144 is described in greater detail throughout this disclosure.

In embodiments, data processing system 12142 may include data processing resources that may be centralized and/or distributed, and may include general purpose chipsets, specialty chipsets, and/or configurable chipsets. Data processing system 12142 may include one or more processors that provide scalable computational capabilities for robot control system 12150 including various intelligence resources within intelligence layer 12140. A processor within data processing system 12142 may communicate with a number of peripheral devices via a bus system. Peripheral devices include, for example, a memory subsystem for storage of instructions and data and a file storage subsystem that provides persistent storage for program and data files, a network interface system that provides an interface to an external network, data allocation, A data management system with capabilities including data caching, data pruning and data management and access and control to intelligence and data resources, and a data store including user interface input and output devices.

In an embodiment, data processing system 12142 includes data handling services 12166 and data processing services 12167. Data handling service 12166 is configured to store, retrieve, and otherwise manage data in MPR 12100. In embodiments, data handling service 12166 accesses a set of data stores 12168 and/or libraries 12169, whereby data handling service 12166 stores data on behalf of other components of MPR 12100. Write and read data from 12168 and/or library 12169. In embodiments, data processing services 12167 perform data processing operations on behalf of various components of MPR 12100. For example, the data processing service 12167 may perform database operations (eg, table join, search, etc.), data fusion operations, etc.

In an embodiment, module management system 12148 coordinates the use and configuration of various control interface modules 12130 and physical interface modules 12122, as described below.

In embodiments, communication system 12152 may provide efficient high-speed electronic and wireless communication between components and subsystems of MPR 12100, as well as a fleet operating system and its elements as described herein, external data sources 12036, It is configured to enable communication of the MPR 12100 with third party systems (e.g., via the Internet, etc.), robotic operating organization units, support systems and equipment, human fleet resources, etc. The communication system 12152 is one such as wired, wireless, etc. capable of supporting various data protocols such as Internet Protocol (IP), Bluetooth communication protocol, wireless communication protocol (e.g., IEEE 802, 4G communication protocol, 5G communication protocol), etc. It can include or provide access to more than one type of network. In embodiments, communication system 12152 may utilize intelligence services to organize, prioritize, and control data and resources for various systems within and external to MPR 12100.

In an embodiment, navigation system 12154 allows MPR 12100 to establish its own position and orientation within the environment (localization) while creating a map of the environment as it moves within the environment (localization). Allows navigation through unknown environments. In some embodiments, navigation system 12154 may utilize Simultaneous Localization and Mapping (SLAM) for autonomous navigation of the robot by using sensors to recognize its own location while mapping the environment. The SLAM algorithm generates a map of the surrounding environment from an initial location and estimates the location of the robot and a map of the surrounding environment by repeating the process of discovering the location of the moved robot based on the generated map. Navigation system 12154 may also utilize additional or alternative navigation algorithms.

In embodiments, navigation system 12154 may operate with vision and sensing system 12112 to generate one or more images of MPR 12100 within its environment. These images may be clicked by the cameras and image sensors of vision and sensing system 12112 and may include one or more images clicked using camera 12608 with conformal variable focus liquid lens 12612. . The image may be for a machine vision system 12618 and may utilize one or more neural network models, including a CNN or RCNN, to locate the MPR 12100. Additionally, a number of other sensors, such as motion sensors, depth sensors, proximity sensors, LIDAR, etc., may be used in conjunction with each other to more accurately localize MPR 12100 within the environment.

Additionally, in some embodiments, navigation system 12154 may incrementally build and/or update a map of the environment, where the “map” represents a field of static objects surrounding the robot. MPR 12100 traverses through this map and attempts to measure the range for each object through imaging, laser range finding, or ultrasound, both the position of the detected object and its own position relative to the object. is updated continuously.

In embodiments, the navigation system 12154 may also use motion control to plan a path for the robot and/or task management system 12144 (together with the robot-level intelligence layer 12140) to determine an optimal navigation policy within the environment. Can work with planning system (12158). In some embodiments, navigation system 12154 coordinates with robot control system 12150 to provide control instructions to execute movement of one or more actuators or motors according to a navigation policy that enables MPR 12100 to navigate the environment. creates .

In an embodiment, safety and compliance system 12156 is configured to perform safety assessments including mechanical safety, electrical safety, and functional safety. In embodiments, safety and compliance system 12156 is configured to ensure compliance with one or more safety standards, generate workflow and process control documentation, and obtain a certificate of conformity from one or more standards or certification bodies. In embodiments, safety and compliance system 12156 ensures compliance with one or more standards organizations, including the International Organization for Standardization (ISO), Underwriters Laboratories (UL), TUV SUD, American National Standards Institute (ANSI), etc. . For example, ISO 10218 describes four distinct robot-human collaborative operation modes to ensure that humans are not exposed to unacceptable risks. Similarly, ISO/TS 15066 provides technical specifications and engineering guidance for users to perform risk assessments when installing collaborative robots. In some embodiments, safety and compliance system 12156 may utilize intelligence services when performing safety assessments.

In an embodiment, motion planning system 12158 may be configured to control the motion of MPR 12100 or portions thereof and establish an optimal collision-free path to MPR 2B00. An example embodiment of motion planning system 12158 is described in more detail in conjunction with FIG. 140.

In embodiments, controller 12160 within the control system may drive an actuator, end effector, or any other electro-mechanical component of MPR 2B00 within transportation system 12110, thereby causing MPR 12100 to Enables at least part of a task to be performed. In an embodiment, controller 12160 receives signals from one or more of navigation system 12154, task management system 12144, motion planning system 12158, communication system 12152, and/or module management system 12148. Receive and determine control signals to issue to the associated actuator, which controller 12160 can output to the associated actuator.

In embodiments, modular system 12120 may be configured to provide one or more task-specific capabilities to MPR 12100 using one or more configurable and interchangeable hardware and software modules. In an embodiment, module system 12120 includes control interface module 12130 and/or physical interface module 12122. In embodiments, control interface module 12130 may include one or more software modules to provide connectivity, power, security, sensing, computing, and artificial intelligence (AI)-like capabilities. In embodiments, physical interface module 12122 may include one or more end effectors, or end-of-arm tooling systems, configured to provide MPR 12100 with the ability to perform specific operational tasks.

In an embodiment, control interface module 12130 may include one or more interfaces configured to receive respective modules configured to enhance various capabilities of MPR 2B00, such as sensing capabilities, power capabilities, networking capabilities, edge computing capabilities, etc. Includes. These capabilities may enable MPR (2B00) to perform specialized functions such as specialized sensing and evaluation and to operate in environments with edge and networking constraints, power constraints, mobility constraints, etc.

In an embodiment, control interface module 12130 includes networking module 12131, sensor module 12132, computing module 12133, security module 12134, AI module 12135, communication module 12136, and user interface module. May include (12138). In an embodiment, control interface module 12130 receives one or more sensor modules 12132. The sensor modules used to configure MPR 12100 may depend on the tasks and operations that MPR 12100 is configured to perform. For example, sensor module 12132 may include weight sensors, environmental sensors (e.g., temperature, humidity, ambient light, motion sensors, vision sensors (e.g., cameras, lidar sensors, radar sensors, etc.), or other Suitable sensors may be included. In embodiments, sensor module 12130 may be packaged in a lab-on-a-chip package, an organ-on-chip package, or an organ-on-chip package. ), etc. may be specialized chips.

In an embodiment, control interface module 12130 includes one or more modular, removable and replaceable lab-on-a-chip sensor packages to provide chemical and biological sensing. The lab-on-a-chip sensor package can enable the MPR (2B00) to perform chemical and diagnostic testing, including chemical assays, microbiological culture assays, immunoassays and nucleic acid assays, environmental condition testing, water and gaseous particle analysis, first responder testing, toxicology, military, disaster, and related applications.

In an embodiment, control interface module 12130 includes one or more modular, removable and replaceable organ-on-chip sensor packages tailored to detect and assess biological and related hazards. Organ-on-chip sensor packages can be microfluidic culture devices that simulate the architecture, dynamics, function, and physiological responses of living human organs, including lung, intestine, kidney, skin, bone marrow, and blood-brain barrier, among others. Some example use cases include first responder, operator health, pandemic, and related applications.

In embodiments, control interface module 12130 may be one or more modular, resettable, and replaceable modules configured to detect potential collisions and disengage or send signals to the robot to stop or reverse movement when a collision is detected. Available includes crash sensor package. Crash sensor packages can help prevent, reduce or eliminate damage to end effectors, tooling, and the parts or products being processed.

In embodiments, control interface module 12130 may be configured for one or more modular, removable, and replaceable AI-on-a-chip (AI) components configured for specific tasks or policies and integrated to operate with a variety of visual and other sensor inputs. -on-a-chip) package. Some examples of task-specific AI-on-a-chip packages include a machine vision package, a natural language processing package, an image classification package, a video analytics package, a predictive analytics package, an optimization package, a control package, or a policy library configured to implement one or more policies. Includes package. In embodiments, a modular AI-on-a-chip package may be configured for training one or more of a machine learning model, a reinforcement learning model, a neural network, a policy network, etc. In embodiments, modular AI-on-a-chip packages may be configured for specific environments, such as warehouses, manufacturing environments, farming and agriculture environments, shipping and logistics environments, healthcare environments, etc. The modular AI-on-a-chip package can be trained with domain-specific models built for specific environments or use cases. For example, a package may include a natural language processing model specifically tailored to understand language used in agricultural or warehouse environments. As another example, a model can be trained on a set of medical images and used to identify microbial infections. In embodiments, control interface module 12130 may be comprised of one or more modular, removable and replaceable AI-configurable devices configured for specific environments, including environments with low or intermittent power, extreme environmental conditions, high temperatures and low heat dissipation, etc. Includes on-a-chip package. In embodiments, a modular AI-on-a-chip package may be configured to autonomously optimize local resources based on task-specific requirements, including: optimization for computation; Storage; network; energy; Heating/cooling capacity; battery capacity; personnel skills; space; Additive manufacturing capabilities, etc.

In an embodiment, a modular AI-on-a-chip package can be used to recognize situations (bottlenecks in a warehouse, congestion/lines in a store, low/sparse customer mix in a part of the environment), object Classifying and recognizing /faces/products/emotions, setting demand-side parameters (price, promotion, advertising position); It can be trained as a model to execute and govern robotic process automation, such as managing supply-side interactions, including managing onboard chatbot interactions, managing recommendation engines to recommend baskets of complementary products, etc. . In embodiments, a modular AI-on-a-chip package may be trained with a model to analyze a user's physiological, neurological, emotional, and cognitive states and customize the response of MPR 12100 based on those states. You can. For example, the package analyzes the user's facial expressions, dialogue, tone, and body movements to determine state, analyzes state information to derive information about customer interests, responses, preferences, etc., and then uses this information to deliver content, It can be supplied to edge devices for product recommendations, advertising, etc. In embodiments, the modular AI-on-a-chip package may be trained as a model to analyze security threat vectors and other vulnerabilities for the MPR 12100 or robot fleet. For example, the package may include biometric analysis, behavioral modeling, face and voice recognition to enable authentication; Learning models to recognize and prevent attacks by malware, spyware, ransomware, viruses, worms, Trojans, etc.; You can use classification, clustering, or regression models for threat intelligence, anomaly detection, network and end-point security, and more. In an embodiment, a modular AI-on-a-chip package determines seed and crop selection and optimizes the use of agricultural resources, including land, water, and nutrients, to determine weather conditions, light, and temperature collected from farms in agricultural planning. , can be trained as a model to analyze water use or soil conditions. MPR 12100 can, for example, use the information to follow planting and nutrition routines, perform phenotypic analysis for selective breeding, and use AI-controlled LED light to provide optimized light wavelengths for crops. . In embodiments, a modular AI-on-a-chip package can be trained as a model to detect diseases, pests, weeds, and nutrient deficiencies in soil or crops on an agricultural farm. For example, the MPR (2B00) uses the transport system's propeller or a miniaturized jet engine to fly over a farm, uses the vision and sensing system's cameras to capture images of the farm, and then uses a modular AI-on-Air. -Chip packages can be used to identify problem areas and potential improvements. For example, the image may show the presence of unwanted plants or weeds. MPR 12100 may then make decisions regarding treatment with herbicides or select one or more end effectors to eliminate weeds. In embodiments, a modular AI-on-a-chip package can be trained as a model for monitoring and harvesting crops, plants, fruits, and vegetables of various shapes and sizes. For example, the package could use machine vision and other sensors to identify crops that are ready to be harvested. The package can also navigate the farm, estimate the position and orientation of crops relative to the MPR (12100), identify fruits and vegetables of different shapes and sizes, select suitable end-effectors for selective harvesting, and finally May include disciplined policies for storing or packaging fruits and vegetables. In an embodiment, a modular AI-on-a-chip package can be trained with a model to manage a controlled closed-loop environment for an aquaponics system based on the requirements of plants and fish. For example, an example modular AI-on-a-chip package can receive sensed oxygen levels in an aquatic environment and determine whether the water is sufficiently oxygenated, under-oxygenated, or over-oxygenated. In embodiments, a modular AI-on-a-chip package may be trained with a model to optimize 3D printing parameters.

In embodiments, control interface module 12130 may receive multiple modular, removable, and replaceable combinations of modules to perform specific tasks. For example, in some embodiments, control interface module 12130 may receive lab-on-a-chip capabilities for gas detection and AI-on-a-chip capabilities for machine vision. MPR 12100 may use such a package, for example, for gas leak detection and isolation in above-ground and underground gas pipelines. In this example, MPR 12100 can move along a pipeline and analyze gas concentrations very close to potential leak points. In determining a gas leak, the MPR 12100 can use cameras and IR sensors to click images, use machine vision capabilities to find the leak, and use a policy library to identify one or more policies to correct the leak. You can.

In embodiments, physical module interface 12122 receives (or otherwise connects to) physical actions that may be taken by MPR 12100 and/or auxiliary physical modules that change the physical operation of MPR 12100. Some examples of physical module interfaces 12122 include end effectors 12124, activation adapters 12126, 3D printer adapters 12128, etc. End effector 12124 includes a device or tool that can be connected to the end of an arm of MPR 12100 to manipulate an object or accomplish one or more tasks. For example, different end effectors may be used to grip and grip, lift, place, palletize, brush, drill, inspect, and/or test objects. MPR 12100 may be comprised of one or more end effectors, and thus one or more end effectors may be selected based on a number of factors including: the task(s) to be performed; The size, shape, surface and weight of the object to be manipulated; the object's environment, including the available material clearance around the object; available power source; The precision or accuracy required by the task; etc. It is understood that the end effectors used by MPR 12100 may be selected by fleet management platform 12000 during configuration and/or by MPR 12100 during deployment.

In some example embodiments, the end effector may include a gripper for grasping and gripping objects for a wide range of material handling applications, from stacking large boxes to handling small delicate electronic components. In some example embodiments, fingers or jaws may be attached to the gripper to grasp or hold objects as well as to pick up and place them on, for example, an assembly line, conveyor system, or other automated system. there is. For example, a parallel gripper may have two fingers arranged parallel to each other that can be closed on an object to hold and grip the object, while an inclined gripper may have a variety of different fingers, such as three fingers offset by 120°. The suction gripper may have fingers of angular openings, and the suction gripper may have one or more suction cups for engaging the surface of the object and using negative or suction pressure or vacuum to grip the object; Electromagnetic grippers can be used to grip metal objects, hydraulic grippers driven by hydraulic fluid can be used for high load applications such as lifting heavy objects, and soft grippers can be used for gripping different shapes and sizes such as fresh fruits and vegetables. can mimic human fingers to pick and manipulate delicate objects, Bernoulli grippers can use airflow to attach to objects without physical contact, can be used to handle sterile materials to prevent contamination, and so on. In embodiments, the gripper may include sensors that assist the gripper in finding, handling, and positioning the product. In embodiments, the gripper may include accessories such as force torque sensors and compliant force feedback systems for force control processes that require precise application of force. In embodiments, the gripper may be driven by compressed air, vacuum, or electricity. In some example embodiments, end effector 12124 may be used for applications such as arc welding, spot welding, paint spraying, machining, drilling, water-jet cutting ( A variety of processes including water-jet cutting, flaming, riveting, grinding, deburring, assembling, additive manufacturing, injection molding, etc. You can have a wide variety of process tooling devices attached for your application.

In embodiments, mobility adapter 12126 may include suitable modular components that allow MPR 12100 to traverse specific environments and/or conditions. For example, maneuvering adapter 12126 may include a different set of wheels, moveable legs, fins, jets, turbines, or other suitable vehicles.

In an embodiment, 3D printer adapter 12128 integrates a set of integrated additive manufacturing capabilities for printing as needed. For example, additive manufacturing capabilities may include printing tools, such as agricultural tools or parts, construction tools or parts, packaging tools or parts, replacement parts, and/or other suitable additive manufacturing that allows robots to print items as needed. May include abilities. In this embodiment, the additive manufacturing capability may include appropriately dimensioned printing devices for printing items, as well as any materials needed for printing.

The above description of the different modules is provided for each type of physical module and control module, for example. It is understood that physical module interface 12122 and control module interface 12130 may receive other additional or alternative modules without departing from the scope of this disclosure.

140 is an example architecture of a robot control system 12150 showing details of various components of the robot control system, according to some embodiments of the present disclosure. In an embodiment, intelligence layer 12140 receives requests from a set of intelligence layer clients and responds to these requests by providing intelligence services (e.g., decisions, classifications, predictions, etc.) to these clients. At the robot level, such clients may be part of the robot control system 12150, including the performance management system 12146, task management system 12144, module management system 12148, navigation system 12154, motion planning system 12158, etc. various components and subsystems; Energy storage and power distribution systems (12104), electromechanical and electrofluidic systems (12108), transportation systems (12110), vision and sensing systems (12112), and structural systems (12114) or modular systems (12120) or robotic security systems ( It may include various components of the baseline system 12102, including other suitable systems of the MPR 12100, including 12170).

As an example, intelligence layer 12140 may take as input sensor data including environmental data, video camera streams, maps, audio streams, images, coordinates of known obstacles, etc. from vision and sensing system 12112. Intelligence layer 12140 then coordinates with motion planning system 12158 to make one or more decisions regarding the motion of MPR 12100 or portions thereof, and with navigation system 12154 to make decisions regarding navigation in the environment. May coordinate with task management system 12144 to coordinate and make decisions regarding performing one or more tasks. Controller 12160 within robot control system 12150 may then generate control instructions to drive actuators that enable MPR 12100 to move, navigate in the environment, and perform various tasks.

In embodiments, motion planning system (MPS) 12158 may be configured to control the motion of MPR 12100 or portions thereof (e.g., end effector, end of arm tool). In an embodiment, the motion planning system (MPS) 12158 may specify a series of transitions along which the MPR 12100 can navigate from a “start state” to a “goal state” without colliding with any obstacles in the environment. You can. In embodiments, the starting state and goal state may be determined based on the task or sub-task to be performed. The starting state and goal state can be expressed as the location of the robot, the posture of the robot, the geographic location of the robot, etc.

In some embodiments, MPS 12158 may send one or more images and other sensor data from vision and sensing systems, as well as information representing “starting states” and “target states” (e.g., navigation system 12154 or other can be taken as input (from an appropriate component). In an embodiment, MPS 12158 may then build a motion plan for the robot. In some embodiments, the motion plan is a motion plan graph that represents the geometry of the environment with the states of MPR 12100 as nodes and transitions between states as edges of the graph. In an embodiment, a graph search may be performed to find a path between nodes representing a “start state” and a “goal state.” MPS 12158 may also perform a collision evaluation to determine the collision probability between the MPR 12100 and one or more obstacles in the path and assign cost values to edges of the graph based on the collision probability for the corresponding transition. MPS 12158 may perform minimum cost analysis on the motion plan graph to determine the set of transitions or paths from the “start state” to the “goal state”. In an embodiment, MPS 12158 may coordinate with intelligence layer 12140 and navigation system 12154 to implement a navigation policy with an identified set of transitions or paths. MPS 12158 may also coordinate with controller 12160 to generate control instructions to operate one or more actuators or motors within MPR 12100 to execute a motion plan.

In an embodiment, MPS 12158 considers various kinematic, geometric, physical, and temporal constraints, as well as additional constraints including complex tasks (e.g., manipulation of objects) and uncertainties (movement of one or more obstacles). It can be configured to identify an optimal collision-free path in a 3D workspace while taking into account Collision detection determines whether a volume in 3D space swept by the MPR 12100 moving from one state to another collides with any obstacles. The surfaces and obstacles of the swept volume can be represented as polygons, and collision detection involves computing whether these polygons intersect.

In embodiments, MPS 12158 may utilize one or more machine learning models 12664 within intelligence layer 12140 to adapt the motion plan to real-time changes in the environment. For example, the motion plan may be adapted based on changes in the task performed by the MPR 12100, changes in the end effector 12124, etc. In an embodiment, MPS 12158 may improve its motion planning efficiency by using transfer learning to leverage learning from one task to a related task.

In embodiments, MPS 12158 may receive sensor data from one or more sensors of vision and sensing system 12112 to determine any moving obstacles and movement in the environment based on a machine learning model 12664. One or more machine learning models 12664 may be utilized to predict the trajectory of the obstacle. The MPS (12158) uses the predicted trajectory information to calculate a cost function while considering the probability and cost of collision with a moving obstacle.

In embodiments, MPS 12158 may utilize a 3D path planning algorithm to determine the optimal path. For example, a sampling-based algorithm can use information from a graph consisting of randomly sampled nodes and connected edges in a given configuration space to determine feasible paths for the robot's motion. This random approach has strong advantages in terms of quickly providing solutions to complex problems, such as in high-dimensional configuration spaces. Examples of 3D path planning algorithms that can be used by an MPS system include Visibility Graphs, random search algorithms such as fast traversal of random trees, Probabilistic Road Maps (such as Dijkstra's Algorithm, A ^* Algorithm) Includes optimal search algorithm and bioinspired planning algorithm.

In an embodiment, navigation system 12154 uses the route determined by MPS 12158 (e.g., optimal path) is used. In some embodiments, navigation system 12154 may be used by robot control system 12150 to generate control instructions to execute movement of one or more actuators or motors according to a navigation strategy that allows MPR 12100 to navigate the environment. and adjust. The navigation actions of MPR 12100 may be evaluated by reinforcement learning system 12668 in an iterative manner to continuously update the navigation policy.

In an embodiment, task management system 12144 includes one or more of task execution system 12022, library 12314, vision and sensing system 12112, and robot-level intelligence layer 12140 of fleet operating system 12002. Coordinate between services to execute tasks. In some example embodiments, task management system 12144 may be adapted to complete a task upon receiving a task request (e.g., from a user, from a fleet management platform, and/or from another robot). You can refer to the policy library to identify more pre-trained policies. For example, upon receiving a request to move an object from one place to another, task management system 12144 may identify a holding policy and a navigation policy to complete the task. Task management system 12144 also operates in conjunction with vision and sensing system 12112 to analyze visual and sensor information and past motion history to evaluate one or more objects that may be used in a task and perform assigned tasks for those objects. You can determine one or more actions needed to do this. For the example task of moving an object, the problem of grasping the object may be more complex when there is no past action history or policy and the MPR 12100 encounters the object for the first time (e.g., when it has not encountered it during training). You can. Moreover, the correct gripping technique may differ based on object characteristics. For example, the point at which an object is grasped and the force that can be applied during gripping may be very different for different objects (e.g., depending on consistency, fragility, shape, size, etc.). MPR 12100 may need to work with a wide variety of objects having different shapes or forms, such as glasses, boxes, boxes with side handles, markers, flower pots, manufacturing parts, machine tools, desks, chairs, lamps, etc. , may require different techniques and accessories to grasp and pick up these objects. Task management system 12144 may utilize intelligence layer 12140 to identify object characteristics and adapt policies based on these characteristics. For example, the force applied while grasping an object can be adjusted based on whether the object is made of a delicate material such as glass or ceramic as opposed to if the object is made of metal. As another example, MPR 12100 may use side handles to grip a box when such handles are available. Accordingly, task management system 12144 may also work with module management system 12148 to identify and select appropriate end effectors 12124 or other accessories needed to complete the task.

In embodiments, when a suitable end effector is not found, task management system 12144 may use intelligence layer 12140 (e.g., , machine learning services, RPA services, etc.), which can subsequently be ordered or printed. In the latter scenario, task management system 12144 can leverage the additive manufacturing system and associated design advantages to print suitable end effectors that meet task requirements and specifications as defined by task management system 12144. there is.

In embodiments, task management system 12144 may include one or more policy libraries that define sets of pre-trained policies for performing common robotic tasks. A policy is simply a sequence of actions that need to be taken by the MPR 12100 to perform a task. Some examples of common tasks for which policies may be provided are: navigation, gripping, lifting, transport, counting, sorting, stacking, cleaning, twisting, bending, pressing, drilling, polishing, loading/unloading, assembly/disassembly, Includes packaging/unpackaging, palletizing/depalletizing, grinding, welding, painting, sealing, planting, harvesting, cutting, pruning, weeding, etc. In embodiments, a policy library may include multiple additional or nested learning loops for complex or multi-step tasks. For example, transporting an object from a source to a destination may involve grasping and lifting the object, and then navigating to the destination and placing the object there. In embodiments, the policy library may reference task definitions available in library 12314 to ensure consistency with overall task allocation.

Policies may be defined and updated in any suitable way. In some embodiments, policies may be defined by human users (eg, programmers). In some embodiments, task management system 12144 includes an intelligence layer to learn and optimize policies based on the quality of task completion (quality may be measured by metrics such as destruction, task completion rate, safety, accuracy, etc.). (12140) (e.g., RPA service). In some embodiments, policies may be pre-trained using training data collected from expert demonstrations. For example, training data for welding may be obtained from a professional welder engaged in the act of welding. Data may be obtained from a real-world setting, such as a manufacturing shop, or from a controlled environment. In some embodiments, policies may be pre-trained using training data collected from a simulation environment. For example, training data for grasping can be acquired using a digital twin system that performs simulations using one or more of the arms and end effectors.

In embodiments, policies may be pre-trained for a wide variety of objects and adapted based on the characteristics of the objects to which they apply. For example, to train a grip policy, the digital twin system 12630 performs simulations on different objects including glasses, a box, a box with side handles, a marker, a flower pot, a manufacturing part, a machine tool, a desk, and a chair. It can be done. Additionally, transfer learning can be used to adapt or tune data collected for one task to another related task. For example, transfer learning can reuse a model developed on one task as a starting point for a model on a second related task.

In some embodiments, intelligence layer 12140 may use transfer learning for domain adaptation. For example, one or more transfer learning algorithms may be used to adapt data collected by a digital twin system in a simulated environment to a real-world environment. In an embodiment, intelligence layer 12140 may use adversarial training for domain adaptation. For example, generative adversarial networks (GANs) can be used to generate synthetic data about real-world environments, which are then used for training. Additionally, specialized neural networks, such as the Domain-Adversarial Neural Network (DANN) by Ganin et al., can be used for domain adaptation.

ROBOT-LEVEL INTELLIGENCE LAYER

In embodiments, the robot-level intelligence layer 12140 of MPR 12100 may be configured as part of a broader intelligence system (e.g., intelligent service system 12200 of FIG. 130) as described above. In an embodiment, robot-level intelligence layer 12140 provides intelligence services to MPR 12100, allowing MPR 12100 to make decisions, predict, classify, etc. In embodiments, robot-level intelligence layer 12140 may include the ability to perform some or all of the intelligence services consumed by MPR 12100 and/or external sources (e.g., other robots, edge devices). , and/or a fleet management platform).

In an embodiment, intelligence layer 12140 may include an intelligence layer controller 12141 and a set of artificial intelligence (AI) services 12143. In an embodiment, intelligent layer controller 12141 may include an analytics management module 12600, a set of analytics modules 12610, and a governance library 12620. In embodiments, analytics management module 12600 receives requests for artificial intelligence services and determines governance standards and/or analytics associated with the requests. In embodiments, analysis management module 12600 may determine governance standards to apply to a request based on the type of decision requested and/or whether a particular analysis should be performed for the requested decision. For example, a request for a control decision that would cause MPR 12100 to navigate to a nuclear waste disposal site could be tied to a specific set of applicable governance standards, such as safety standards, legal standards, quality standards, regulatory standards, financial standards, etc. and/or may involve one or more analyzes related to control decisions, such as hazard analysis, safety analysis, engineering analysis, etc. In embodiments, governance standards may be defined as a set of standards libraries stored in governance library 12620. In embodiments, governance library 12620 may define conditions, thresholds, rules, recommendations, or other appropriate parameters under which decisions can be analyzed. In some embodiments, analytics management module 12600 may determine one or more analyzes to be performed in connection with a particular decision and provide artificial intelligence service 12143 with a corresponding analytics module 12610 to perform the analyses. there is. In embodiments, analysis module 12610 may include modules configured to perform specific analyzes on specific types of decisions, where each module hosts an instance of intelligence layer 12140: a data processing system ( 12142). Continuing with the example of a decision for MPR 12100 to navigate to a nuclear waste disposal site, hazards and risk levels at the site may need to be analyzed to make navigation decisions. Non-limiting examples of analysis modules 12610 include risk analysis module(s), security analysis module(s), decision tree analysis module(s), ethics analysis module(s), failure mode and effects (FMEA) analysis module ( s), risk analysis module(s), quality analysis module(s), safety analysis module(s), legal analysis module(s), financial analysis module(s), and/or other suitable analysis module(s).

Artificial intelligence services (12143) include digital twin systems (12630), machine vision systems (12618), machine learning (ML) systems (12632), robotic process automation (RPA) systems (12652), and natural language processing (NLP) systems (12656). ), an analysis system 12660, and/or a neural network system 12662. Machine learning system 12632 may further include a machine learning model 12664 and a reinforcement learning system 12668.

The digital twin system (12630) is a digital twin for the MPR (12100), robotic subsystems such as electromechanical and electrofluidic systems (12108), transportation systems (12110), vision and sensing systems (12112), batteries, sensors, valves, etc. , robot components such as actuators, motors, end effectors, etc., and can be configured to create robot policies such as navigation, grasping, lifting, transport, etc. The digital twin of the MPR 12100 may have a visual user interface, for example in the form of a 3D model, and/or may consist of a system specification or ontology that describes the architecture, including the components and interfaces of the MPR 12100. You can. The digital twin can be configured to simulate the operation of the MPR 12100 to continuously capture key operational metrics and can be used to monitor and optimize the performance of the MPR 12100 in real time. A robot digital twin may also be configured to communicate with one or more users, twins, or other robots through multiple communication channels such as conversation, text, gestures, etc. For example, the digital twin may receive a query regarding MPR 12100 from a user, generate a response to the query, and communicate such response. Additionally, the digital twin system 12630 may be configured with an interface such as an API for receiving information from the operating environment of the MPR 12100.

In embodiments, digital twin system 12630 may be used to simulate the behavior of MPR 12100 or one or more of its components or subsystems. For example, the behavior of MPR 12100 while grasping a glass bottle and moving it from a source to a destination may be predicted and optimized by intelligence layer 12140. Insights gained from analytics and simulations using digital twins can be passed on to reinforcement learning agents to improve these processes.

In embodiments, multiple digital twins of components and subsystems of MPR 12100 may be integrated, thereby aggregating data across the value chain network to create a digital twin for MPR 12100 and provide entity-level insights. In addition, it can also drive system-level insights. Similarly, digital twins of policies can be combined to form digital twins of multi-step tasks or task twins. For example, a digital twin for transportation could appear to consist of digital twins of gripping, lifting and navigation.

Machine vision system 12618 includes software that allows MPR 12100 to extract information from digital images to recognize one or more objects within the environment of MPR 12100. Machine vision system 12618 may execute one or more machine learning algorithms to perform one or more machine vision tasks, including object classification, object detection, scene classification, pose detection, semantic segmentation, instance segmentation, and image captioning. You can. Machine vision systems can include pre-trained machine learning models to execute different machine vision tasks, including neural networks-like, convolutional neural networks (CNNs), transformer neural networks, region-based CNNs, fast RCNNs, mask RCNNs, etc.

MACHINE LEARNING SYSTEM

Machine learning system 12632 may include one or more machines for performing analysis, simulation, decision-making, and predictive analytics related to data processing, data analysis, simulation generation, and simulation analysis of one or more components or subsystems of MPR 12100. A learning model (12664) can be defined. In embodiments, machine learning models 12664 are algorithms and/or statistical models that perform specific tasks without using explicit instructions and instead relying on patterns and inferences. Machine learning models 12664 build one or more mathematical models based on training data to make predictions and/or decisions without being explicitly programmed to perform a specific task. In an example implementation, machine learning model 12664 may perform classification, prediction, regression, clustering, anomaly detection, recommendation generation, and/or other tasks.

In embodiments, machine learning model 12664 may perform various types of classification based on input data. Classification is a predictive modeling problem in which class labels are predicted for given examples of input data. For example, machine learning models can perform binary classification, multi-class, or multi-label classification. In an embodiment, a machine learning model may output “confidence scores” that indicate the respective confidence levels associated with the classification of the input into each class. In embodiments, confidence scores may be compared to one or more thresholds to render discrete category predictions. In an embodiment, only a certain number of classes (e.g., one) with the relatively largest confidence score may be selected to render a discrete category prediction.

In embodiments, machine learning model 12664 may output a probabilistic classification. For example, a machine learning model can predict a probability distribution over a set of classes, given sample input. Therefore, rather than outputting only the most likely class to which the sample input should belong, the machine learning model can output, for each class, the probability that the sample input belongs to that class. In an embodiment, the probability distribution for all possible classes may sum to 1. In embodiments, a softmax function, or other type of function or hierarchy, may be used to convert a set of real values each associated with a possible class into a set of real values in the range (0, 1) that sums to 1. In embodiments, the probabilities provided by the probability distribution may be compared to one or more thresholds to render discrete category predictions. In embodiments, only a certain number of classes (e.g., one) with the relatively greatest prediction probability may be selected to render discrete category predictions.

In embodiments, machine learning model 12664 may perform regression to provide output data in the form of continuous numeric values. By way of example, a machine learning model may perform linear regression, polynomial regression, or non-linear regression. As described, in embodiments, the softmax function or other function or layer squashes the set of real values each associated with two or more possible classes into a set of real values in the range (0, 1) that sums to 1. can be used to For example, machine learning model 12664 may perform linear regression, polynomial regression, or non-linear regression. As an example, machine learning model 12664 may perform simple regression or multiple regression. As described above, in some implementations, the softmax function or other functions or hierarchies convert a set of real values, each associated with two or more possible classes, into a set of real values in the range (0, 1) whose sum is 1. Can be used for quashing.

In embodiments, machine learning model 12664 may perform various types of clustering. For example, a machine learning model may identify one or more previously defined clusters to which the input data is most likely to correspond. In some implementations where the machine learning model performs clustering, the machine learning model may be trained using unsupervised learning techniques.

In embodiments, machine learning model 12664 may perform anomaly detection or outlier detection. For example, a machine learning model may identify input data that does not conform to expected patterns or other characteristics (e.g., as previously observed from previous input data). As an example, anomaly detection may be used to detect fraudulent transactions or detect system failures.

In some implementations, machine learning model 12664 may provide output data in the form of one or more recommendations. For example, machine learning model 12664 may be included in a recommender system or engine. For example, given input data that describes previous outcomes for a particular entity (e.g., scores, rankings, or grades representing amounts of success or enjoyment), a machine learning model can, based on the prior outcomes, produce the desired outcome. It may output suggestions or recommendations of one or more additional entities that are expected to have .

As described above, machine learning model 12664 may be or include one or more of a variety of different types of machine learning models. Examples of these different types of machine learning models are provided below for illustration. One or more of the example models described below may be used (e.g., combined) to provide output data in response to input data. Additional models other than the example models provided below may also be used.

In some implementations, machine learning model 12664 includes one or more classifier models, such as, for example, a linear classification model; It may be or include a secondary classification model, etc. Machine learning models 12664 may include one or more regression models, such as, for example, a simple linear regression model; multiple linear regression model; Logistic regression model; stepwise regression model; Multivariate adaptive regression spline; Locally estimated scatterplot smoothing model; It may be or may include this.

In some examples, machine learning model 12664 can be one or more decision tree-based models, such as, for example, classification and/or regression trees; Chi-square automatic interaction detection decision tree; decision stump; conditional decision tree; It may be or may include this.

Machine learning model 12664 may be or include one or more kernel machines. In some implementations, machine learning model 12664 may be or include one or more support vector machines. Machine learning model 12664 may include one or more instance-based learning models, such as, for example, a learning vector quantization model; Self-organizing map model; Locally weighted learning model; It may be or may include this. In some implementations, the machine learning model includes one or more nearest neighbor models, such as, for example, a k-nearest neighbor classification model; k-nearest neighbor regression model; It may be or may include this. Machine learning model 12664 may include, for example, one or more Bayesian models, such as a naive Bayes model; Gaussian naive Bayes model; multinomial naive Bayes model; averaged 1-dependency estimator; Bayesian network; Bayesian trust network; Hidden Markov Model; It may be or may include this.

In some implementations, machine learning model 12664 may be or include one or more artificial neural networks (also referred to simply as neural networks). A neural network may contain groups of connected nodes, which may also be referred to as neurons or perceptrons. Neural networks can be organized into one or more layers. A neural network containing multiple layers may be referred to as a “deep” network. A deep network may include an input layer, an output layer, and one or more hidden layers located between the input layer and the output layer. Nodes in a neural network can be connected or completely disconnected.

Machine learning model 12664 may be or include one or more feedforward neural networks. In a feedforward network, the connections between nodes do not form cycles. For example, each connection may connect a node from a previous layer to a node from a later layer.

In some cases, machine learning model 12664 may be or include one or more recurrent neural networks. In some cases, at least some of the nodes of a recurrent neural network may form a cycle. Recurrent neural networks can be particularly useful for processing input data that is sequential in nature. In particular, in some cases, a recurrent neural network may transfer or retain information from a previous portion of an input data sequence to a subsequent portion of the input data sequence through the use of recursive or directed recurrent node connections.

In some examples, sequential input data may include time series data (e.g., time or images captured at different times subject to sensor data). For example, recurrent neural networks can analyze sensor data versus time to detect or predict swipe direction, perform handwriting recognition, and more. Sequential input data includes words within sentences (e.g., for natural language processing, conversation detection or processing, etc.); notes in musical composition; sequential actions taken by the user (e.g., to detect or predict sequential application usage); sequential object states; It may be or may include this.

Exemplary recurrent neural networks include long short-term (LSTM) recurrent neural networks; gated circulation unit; Bidirectional recurrent neural network; Continuous-time recurrent neural network; neural history compressor; echo state network; Elman Network; Jordan Network; recursive neural network; Hopfield Network; fully recurrent network; Sequence-to-sequence composition; Includes etc.

In some examples, machine learning model 12664 may be or include one or more acyclic sequence-to-sequence models based on self-attention, such as a transformer neural network. Details of an example transformer neural network can be found at http://papers.nips.cc/paper/7181-attention-is-all-you-need.pdf.

In some implementations, machine learning model 12664 may be or include one or more convolutional neural networks. In some cases, a convolutional neural network may include one or more convolutional layers that perform convolutions on input data using learned filters.

A filter may also be referred to as a kernel. Convolutional neural networks can be particularly useful for vision problems, such as when the input data includes images such as still images or video. However, convolutional neural networks can also be applied for natural language processing.

In some examples, machine learning model 12664 may be or include one or more generative networks, such as, for example, an adversarial generative neural network. Generative networks can be used to generate new data, such as new images or other content.

Machine learning model 12664 may be or include an autoencoder. In some cases, the goal of an autoencoder is to learn a representation (e.g., low-dimensional encoding) for a set of data, typically for the purpose of dimensionality reduction. For example, in some cases, an autoencoder may attempt to encode input data and provide output data that reconstructs the input data from the encoding. Recently, the autoencoder concept has become more widely used to learn generative models of data. In some cases, autoencoders may involve additional losses beyond reconstructing the input data.

Machine learning model 12664 may be, for example, one or more other types of artificial neural networks, such as, for example, a deep Boltzmann machine; deep trust network; Stacked autoencoder; It may be or may include this. Any of the neural networks described in this application may be combined (e.g., stacked) to form more complex networks.

Machine learning model 12664 may include one or more clustering models, such as, for example, a k-means clustering model; k-median clustering model; expectation maximization model; hierarchical clustering model; It may include etc.

In some implementations, machine learning model 12664 may be implemented using one or more dimensionality reduction techniques, such as, for example, principal component analysis; Kernel principal component analysis; Graph-based kernel principal component analysis; principal component regression; Partial least squares regression; Salmon Mapping; multidimensional scaling; projection tracking; linear discriminant analysis; mixed discriminant analysis; second-order discriminant analysis; generalized discriminant analysis; flexible discriminant analysis; auto-encoding; etc. can be performed.

In some implementations, a machine learning model can be implemented using one or more reinforcement learning techniques, such as a Markov decision process; dynamic programming; Q function or Q-learning; value function approach; deep Q-network; differentiable neural computer; A3C (asynchronous advantage actor-critics); deterministic policy gradient; etc. can be performed or these can be applied.

Reinforcement learning is a machine learning technique for learning optimal behavior in an environment by taking actions and getting feedback, similar to how humans and animals learn by interacting with their environment. A typical reinforcement learning approach observes the environment, evaluates the current state (e.g., robot speed, distance to an object ahead), and selects an action (e.g., provides control commands to an actuator or motor, an agent (e.g., robot control system 12150) that adjusts speed, changes direction, etc. When performing an action, the agent receives, in addition to the new state, a reward that provides some indication of the success of the action (e.g. +10 for allowing sufficient space between the robot and the obstacle in front of it and +10 for allowing insufficient space). You are given -10) for acceptance. The goal of a reinforcement learning agent is to learn the optimal policy or behavior that maximizes the expected cumulative reward.

Reinforcement learning system 12668 includes one or more reinforcement learning algorithms for evaluating various states, actions, and rewards when determining an optimal policy for executing one or more tasks by MPR 12100.

RPA system 12652 enables MPR 12100 to automate workflows as well as any repetitive tasks and processes. In embodiments, RPA system 12652 may monitor human interactions with various systems to learn patterns and processes performed by humans in the performance of each task. In embodiments, RPA system 12652 may learn to perform specific tasks based on learned patterns and processes, such that tasks may be performed by RPA system 12652 instead of or in support of a human decision maker. .

NLP system 12656 provides MPR 12100 with the ability to communicate with a human user as well as parse one or more conversational voice commands provided by the human user to perform one or more tasks. In embodiments, NLP system 12656 may be configured as part of, utilize, or be included in NLP system 4D24 described in conjunction with FIG. 4. The NLP system 12656 is a neural network system 12662 including a feedforward neural network, convolutional neural network (CNN), recurrent neural network (RNN), long short-term memory (LSTM), and transformer neural network to perform various natural language processing functions. More than one neural network can be utilized. Example implementations of NLP system 12656 are described in greater detail elsewhere in this disclosure (e.g., in connection with FIG. 104 and the related description).

In embodiments, artificial intelligence service 12143 may include and/or provide access to analytics system 12660. In embodiments, analysis system 12660 is configured to perform various analysis processes on data output from MPR 12100 or one or more components or subsystems. For example, the analysis system 12660 may perform anomaly detection, system failure detection, predictive maintenance, and monitor MPR(2B00) over a period of time to avoid costly downtime and interruption of operation of MPR(2B00). Data analysis can be performed on heat and vibration data generated by. In another example, analysis system 12660 may analyze sensor data of MPR 12100 to determine the general health of MPR 12100 efficiency of one or more tasks performed by MPR 12100 and the optimal performance for MPR 12100. It can generate insights about things like location and settings.

Neural networks (or artificial neural networks) are a family of statistical learning models that derive their inspiration from biological neural networks, can rely on large numbers of inputs, and are commonly used to estimate or approximate unknown functions. A neural network represents a system of interconnected “neurons” that send messages to each other. Connections have numerical weights that can be tuned based on experience, allowing the neural network to adapt to input and learn.

Neural network system 12662 includes feedforward neural networks, convolutional neural networks (CNN), recurrent neural networks (RNN), long short-term memory (LSTM) neural networks, gated recurrent unit (GRU) neural networks, and self-organizing map (SOM) neural networks (e.g. For example, Kohonen self-organizing neural networks), autoencoder (AE) neural networks, encoder-decoder neural networks, modular neural networks, or variants, hybrids or combinations of the foregoing, or reinforcement learning (RL) systems or other expert systems, such as rules base systems, including one or more neural networks including combinations with model-based systems (including those based on physical models, statistical models, flow-based models, biological models, biomimetic models, etc.). Examples of neural networks and neural network systems 12662 are described in more detail elsewhere in this disclosure (e.g., Figures 93-107).

141 uses data from multiple sensors in the vision and sensing system 12112 to learn about the environment to implement policies and implement policies in the energy storage and power distribution system 12104, electromechanical and electrofluidic system 12108, or Schematically depicts an example architecture of a robot control system 12150 that drives control of one or more components of a baseline system 12102, including a transportation system 12110, to perform tasks.

In embodiments, MPR 12100 may acquire sensor data from one or more sensors 12602 and extract “state information” regarding the location of MPR 12100 relative to the environment 12604 and one or more objects 12606. there is. For example, MPR 12100 may use camera 12608 to capture an image of object 12606. Additional vision sensors may be mounted at locations different from those of camera 12608 to capture image data from multiple viewpoints. Cameras and vision sensors may produce images related to the shape, color, depth, and/or other characteristics of the object(s) in the sensor's field of view. Image data may be processed and machine vision system 12618 may execute one or more machine learning algorithms, including CNN variants described above, for object detection. Data from additional sensors (e.g., tactile sensors, sound sensors, and/or gas sensors) may be combined to help MPR 12100 build a more accurate model of the world to more successfully navigate and maneuver in the environment. You can. In embodiments, Kalman filters and data fusion techniques may be used to combine data from multiple sensors.

In an embodiment, intelligence layer 12140 provides controls to perform one or more tasks including navigation, object grasping, sort cleaning, loading/unloading, packaging/unpackaging, assembly, palletizing/depalletizing, etc. It may coordinate with policy libraries within task management system 12144 and controller 12160 to generate commands.

When control system 12150 receives input (e.g., from a user or from another robot) indicating one or more tasks to be performed, intelligence layer 12140 selects one or more from a policy library within task management system 12144 to implement. You can choose a policy. For example, upon receiving a command to grasp an object placed in the environment of the MPR 12100, the intelligence layer 12140 may determine a navigation policy for the MPR 12100 to navigate to the location of the object, followed by a grab policy to grasp the object. You may decide that you need to use a policy. Intelligence layer 12140 may use sensor data from one or more sensors 12602 to determine “state information” that describes information extracted from the scene within the environment of MPR 12100. State information may include images or image streams from one or more vision sensors, information collected from other sensors such as gas sensors, tactile sensors, and sound sensors. State information may also include information obtained after analysis of sensor information, such as the presence of one or more objects in the environment, the name and type of the object, and a map containing the target object to be grasped for MPR 12100. It may include the distance and position of the object on the image, the material characteristics of the target object, etc.

In an embodiment, intelligence layer 12140 may then take one or more actions based on one or more policies in response to the state information. For example, the intelligence layer 12140 may determine that the environment contains two objects and that the MPR 12100 needs to travel 100 meters to reach the target object while avoiding an obstacle object located at a distance of 10 meters. there is. The navigation policy may provide navigation actions and guide the MPR 12100 to reach the target object while avoiding collisions with obstacle objects. The grasp policy can then guide the MPR 12100 on action steps to grasp the target object. In embodiments, a policy library may use machine learning, including reinforcement learning, to define different policies to perform various tasks.

Based on the output of the policy library in task management system 12144, robot control system 12150 then implements the policy and implements policies in electromechanical system 12108, transportation system 12110, or energy storage and power distribution system 12104. Control instructions may be developed and provided for one or more actuators or control devices associated with the MPR 12100 to drive one or more components of the MPR 12100. For example, control instructions may execute movement of one or more motors of transportation system 12110 to navigate to a location within the environment according to a navigation policy. As another example, control instructions may execute movements in one or more actuators within an arm joint or end effector to grip a target object according to a grip policy.

The term actuator includes a mechanical or electrical device that generates motion, in addition to any driver(s) that may be associated with the actuator and that converts received control commands into one or more signals to drive the actuator. Accordingly, providing control commands to an actuator may include providing control commands to a driver that converts the control commands into appropriate signals for driving an electrical or mechanical device to produce the desired motion. MPR 12100 may have multiple degrees of freedom, and each actuator or motor may control operation within one or more of the degrees of freedom in response to control commands.

142 illustrates an example vision and sensing system 12112 according to some embodiments of the present disclosure. The vision and sensing system 12112 includes a range of sensors 12602 configured to receive information from the environment 12604 of the multipurpose robot 12100 and enable the MPR 12100 to interact with one or more objects 12606 within that environment. Includes. For example, a vision sensor can capture image data within the field of view that can assist the MPR 12100 with environmental awareness and navigation. Some examples of sensors include one or more cameras, LIDAR, RADAR, SONAR, thermal imaging, hyperspectral imaging, light sensor, force sensor, torque sensor, speed sensor, acceleration sensor, position sensor, proximity sensor, gyro sensor, sound sensor, motion sensor, position sensor, load sensor, temperature sensor, touch sensor, depth sensor, ultrasonic range sensor, infrared sensor, chemical sensor, magnetic sensor, inertial sensor, gas sensor, humidity sensor, pressure sensor, viscosity sensor, flow sensor, object sensor , tactile sensors, etc. In embodiments, the sensor may be mounted directly on an inoperable component of the robot, such as a head, or on an actuated component, such as an arm or end effector. In embodiments, sensors may be physically separate from MPR 12100 or located within the environment 12604 in which MPR 12100 is operating.

In embodiments, vision and sensing system 12112 may monitor environment 12604 in real time and detect obstacles, terrain elements, weather conditions, temperature, or other aspects of the environment. The various sensors 12602 are configured to operate in a wide range of environmental conditions and are associated with one or more objects 12606 within the environment 12604, such as the size, shape, profile, structure, speed, distance, or orientation of the objects 12606. Data can be captured. Some examples of sensors 12602 that can operate to capture different data in various environments include monographic cameras (e.g., for capturing image data), stereoscopic cameras (e.g., for 3D vision), ( RADAR (e.g., for long-range object detection, distance determination, or speed determination), LIDAR (e.g., for short-range object detection, distance determination, or speed determination), (e.g., underwater object detection, distance determination) , or SONAR (for speed determination), ultrasonic sensors (e.g. for bright light and very dark environments and for detecting glass or other transparent surfaces), GPS (e.g. for location information), (e.g. For example, it includes an IMU (for orientation information), etc.

In embodiments, vision and sensing system 12112 may coordinate with robot control system 12150 to process captured sensory data and devise policies or make a sequence of decisions regarding actions to be performed by MPR 12100. there is. Determinations may include, for example, activating or deactivating one or more components of the electromechanical and electrofluidic system 12108, movement of the MPR 12100 by the transportation system 12110, and movement of the MPR 12100 by the energy storage and power distribution system 12104. It may be related to the distribution of power to specific components of the MPR (12100).

Referring now to FIG. 142 , camera 12608 is configured to capture an image of object 12606 located within the field of view of camera 12608. Camera 12608 may be a standard digital camera (i.e., a camera containing a CCD or CMOS sensor), a stereoscopic camera, an infrared image sensor, a time-of-flight (TOF) camera, etc., with power/control connections and optical elements such as a lens 12612; It may be a structured light camera, etc. Lens 12612 may be compliant, configured to adjust various optical parameters, including lens shape, focal length, liquid material, reflectivity, color, environment, lens arrangement, for example, via control signals received via a power/control connection. It may be a variable focus liquid lens. In an embodiment, control connections may include electrical, hydraulic, pneumatic, mechanical, thermal or magnetic controls. The compliant liquid lens 12612 may include an autofocus capability to help quickly adjust its focal length, which allows recognizing objects in a dynamic environment, such as when the object 12606 or MPR 12100 moves. ; recognizing three-dimensional (3D) objects by capturing depth data; recognizing small objects; recognizing objects in power-constrained or network-constrained environments; etc. is possible.

Raw image data captured by camera 12608, which can be in various forms including RGB images, thermal images, point clouds, can then be subjected to data transformation, filtering, de-noising, aggregation, artifact reduction, compression, analog-to-digital conversion, It is sent to the preprocessor 12614 to perform data preprocessing, including preliminary feature recognition. The image data is then sent to the image processing engine 12616 for further processing, for example, to identify the object 12606 in the image as well as determine its location or orientation. Image processing engine 12616 may interface with machine vision system 12618 within intelligence layer 12140 of robot control system 12150. Machine vision system 12618 may execute one or more machine learning algorithms to perform one or more machine vision tasks, including object classification, object detection, scene classification, pose detection, semantic segmentation, instance segmentation, and image captioning. You can. A machine vision system may include pre-trained machine learning models to perform different machine vision tasks. In embodiments, machine vision system 12618 may utilize one or more neural network-based models for processing image data.

In an embodiment, the vision and sensing system 12112 may have dynamic vision with artificial intelligence to learn based on a training set of results, parameters, and data collected from the compliant varifocal liquid lens 12612 to recognize objects. Includes system. In embodiments, the dynamic vision system is controlled by and/or optimized by input from an artificial intelligence in the intelligence layer 12140, e.g., the artificial intelligence provides visual information in a manner that improves results, such as recognition results, prediction results, etc. Learn based on a set of machine vision results to adjust the dynamic vision system to capture

In an embodiment, vision and sensing system 12112 includes a dynamic vision system including an optical assembly having a conformable varifocal liquid lens 12616; a robot control system 12150 configured to adjust in real time one or more optical parameters and data collected from the optical assembly; and train a set of machine learning models 12664 based on the training set of results, parameters and data collected from the optical assembly to control the optical assembly to optimize collection of data for processing by the set of machine learning models. Includes a dynamically learning data processing system 12142. In an embodiment, a first model is used to optimize the collection of signals by an optical assembly and a second model is used to act on the signals to achieve a desired machine vision result. In embodiments, the result is a recognition result, a classification result, or a prediction result.

Dynamic vision capabilities provided by vision and sensing system 12112 allow MPR 12100 to identify target objects for use in robotic assembly lines where object depth, orientation, position, and motion can be inferred for improved object identification. and can be manipulated. Dynamic vision capabilities can also enable MPR 12100 to perform simultaneous localization and mapping, a technique for estimating the robot's position relative to its surroundings while simultaneously mapping the environment.

In embodiments, visual output from vision and sensing system 12112 is temporally combined with output from other sensors within MPR 12100 using conditional probabilities to provide richer, more accurate information about the location, orientation, and motion of objects in the environment. A combined view of the target object containing related information can be created.

In embodiments, the dynamic vision capabilities of vision and sensing system 12112 may be integrated into or with a set of Value Chain Network (VCN) entities for quality control inspection and classification of objects on a production assembly line or logistics chain. Here the compliant liquid lens 12612 is configured to quickly adjust focus to accommodate, recognize, and classify objects located at different working distances or at different heights.

104-142, according to some example implementations, a fleet management platform with wireless power routing and management for robotic instrumentation and associated electronics also includes a modular, removable organ-on-chip sensor robot. -on-chip sensor robot) It is easy to configure and operate the robot with sub-assembly. In embodiments, power to the organ-on-chip sub-assembly is provided such that primary power to the robot is provided by a replaceable battery pack and power to the organ-on-chip is provided by a sub-assembly-specific battery pack. It can be delivered and managed wirelessly to satisfy a wide range of robot deployments, including an optional mobile environment. In embodiments, power sharing and routing of power between battery packs may be performed and managed wirelessly, such as by a robot-local power management facility. The platform may facilitate performing fleet configuration based on wireless power routing options available for candidate robots. Examples include, but are not limited to, a single power pack for wirelessly providing power to an on-robot sub-assembly, such as an organ-on-chip sub-assembly driven via a robot-local wireless power routing system. Wireless power routing and management can be extended to removable robotic sensor-type sub-assemblies, such as organ-on-chip examples, that can be placed separately from but within the wireless power routing scope of the robot. This may be useful in environments where sensors and robots cannot be co-located (e.g., due to size, environment, or other constraints). According to some example implementations, a fleet management platform with a control tower for combined control of robots such as MPR, SPR and exoskeleton, and an additive manufacturing system also includes an artificial intelligence system for automated design and 3D printing of robot accessories. You can have it. In some of these examples, artificial intelligence systems can automate design and 3D printing based on contextual task recognition. This task recognition may be based on shape recognition sensors (e.g., vision sensors) and motion history (either about the robot or based on other factors such as the task) to determine, for example, robot end-effector requirements to complete the task. May depend on usage. In embodiments, the results of this AI generated task recognition may be provided to a control tower to further enhance flexible on-demand additive manufacturing based on recognition of the tasks to be performed. In embodiments, the control tower may further combine control of robotic 3D printing for additive manufacturing with robotic control of contextually determined 3D printing of end effectors, thereby adding value to the 3D printing capabilities of the fleet of robots. can be increased. In embodiments, this combination may facilitate field maintenance of robots, production equipment, warranty repairs, etc. Additionally, the use of artificial intelligence to facilitate task recognition can be useful for servicing/repairing production systems in cases where some details of the required task may not be known until the robot participates (e.g. fully automated production operations). Autonomous responsiveness can be improved.

According to some example implementations, a robotic fleet management platform with autonomous local system task allocation adaptability based on sensed local context may also be integrated with supply chain infrastructure entities for improved dynamic supply chain adaptability and efficiency. In some of these example implementations, application of local system task allocation adaptability with supply chain integration may enhance the capabilities of robots deployed within containers, for example. This combination of fleet management capabilities may also facilitate coordination (e.g., based on peer communication, etc.) between robots along the supply chain, e.g., deployed within or alongside smart containers, etc., thereby It can provide flexibility when configuring individual robots in advance. In one example, a set of robots deployed with long-distance trucks, ships, etc. can allocate supply chain tasks among themselves based on locally sensed conditions. The set of tasks to be performed during a transoceanic journey that is part of a supply chain may be adaptively assigned based on local time circumstances, such as local weather conditions.

According to some example implementations, in particular, a robot fleet management platform with smart contract support capabilities for negotiated routing of robots may also have an artificial intelligence-governed data pipeline to support remote robot management. In embodiments, detectable smart contract duration as a function of robot behavior may affect how the AI-governed data pipeline is managed. By way of example, a data pipeline may be configured to, for example, require robots to deliver certain data pipeline requirements (e.g., ensure that high priority data signals are delivered to workers, robots, and/or clients to ensure security, safety, and other concerns). can be managed to ensure that average and peak throughputs are achieved while ensuring that . However, these robotic data pipelines also allow data representing smart contract durations (e.g. timeliness of response, up-time, etc.) to be accurately and timely tracked (optionally recorded, stored, and It can be managed (e.g. through AI-governance) to ensure that it can be delivered later. Within this context, configuring a data pipeline for one or more robots involved in the execution of a smart contract (e.g. to provide assurance services) to the value chain network to update state related to the smart contract terms and conditions. (VCN) may include configuring infrastructure elements. An AI-based data pipeline governance system can, for example, optimize the use of sensor detection packages on robots across the VCN to ensure that data pipeline requirements are met. In one example, a set of robots working collaboratively across a value chain network may have differently configured (e.g., optimized) sensors depending on their relative positions in the value chain network when smart contract terms are taken into account in robot configuration. You can have a package. As another example, the construction and use of an on-robot data store may also allow certain data collected (e.g., through robot sensor packages, etc.) to be stored locally and optionally passed through a data pipeline to a smart contract control facility. It can be influenced by smart contract terminology to be curated/filtered before it becomes available. In this example, data pipeline resources can be prioritized so that only substantial deviations from normal for a particular smart contract term utilize the pipeline. AI-governance of the data pipeline can enable local evaluation of sensed data to influence smart contracts, as long as the information derived from robot behavior regarding satisfying smart contract requirements remains within acceptable bounds. No data pipeline resources are required.

According to some example implementations, the robot fleet management platform with an artificial intelligence (AI) based robot health monitoring system also includes hydraulic flow and actuation systems optimized to reduce hydraulic interconnection through the application of 3D printing in an additive manufacturing environment. You can have In some of these example implementations, the information obtained by the AI-based robot health monitoring system may be used to replace multiple interconnections, for example, with structures with little or no interconnections, for example during preventative maintenance phases. By applying automated design and additive manufacturing, it can be directly applied to mitigate the likelihood of hydraulic interconnect failure. In embodiments, a computer vision system for identifying visual defects or hazards (e.g., identifying hydraulic systems with multiple interconnections), vibration-based detection (e.g., where defect-induced vibration levels are applied), identifying hydraulic interconnection sub-assemblies), multiple interconnection hydraulic system components (e.g. Robot health monitoring systems are provided, such as temperature sensing systems that can provide thermal data regarding (e.g., interconnections, etc.). In embodiments, an AI-based robot health monitoring system can further predict failure areas, such as hydraulic interconnections, that can be used according to additive manufacturing requirements to deliver a potentially more robust hydraulic system. Additionally, failure prediction capabilities can be used as a control over which components should be prioritized for production with an additive manufacturing system. Additionally, the scheduling and routing of robotic systems with additive manufacturing capabilities can be influenced by the predictive capabilities of AI-based robot health monitoring systems, so that services can include the deployment of robotic elements with improved reliability. Service or maintenance by ensuring that additive manufacturing resources are routed to service areas for localized part manufacturing or utilized to produce components (e.g., hydraulic assemblies with fewer interconnections) so that they are locally available when available. The conservative visit value can be optimized.

According to some example implementations, a robotic fleet management platform with artificial intelligence-based shape recognition capabilities for automated task execution may also have a system for coordinated control of robotic systems that incorporates 3D printing for task execution. Robotic sensing and analysis systems analyze visual images and sensor information along with past motion history and task criteria (e.g. definitions, goals, etc.) to evaluate objects associated with a task, such as the object on which the robot motion is to be performed. AI can be used. Object analysis may optionally facilitate determining one or more actions to perform an assigned task, including the type of end effector or other physical interface required to perform the task in a given analysis. In embodiments, one or more operations required may include selection and use of a specific type of end effector, such as a gripper, j-hook, pressure sensitive clamp, gripping and rotation capabilities, etc. In embodiments, 3D printing control capabilities of a robot or companion robot configured to facilitate performing a task are utilized to create suitable end effectors, adapters, or other features based on visual and/or sensed analysis associated with the object. It can be. In an example, an object may have a keyhole type interface for handling the object. Image analysis can be commissioned into the 3D printing control system to detect these features of the object and generate keys suitable for use with the object. Other examples that combine robotic object detection (e.g., shape recognition, etc.) with control of 3D printing capabilities to execute one or more actions of a task associated with the object include non-straight (e.g., circular) objects without an identifiable flat surface. , oval, and rectangle) and detecting the shape of the object. Artificial intelligence-based shape recognition can facilitate detecting the appropriate orientation for lifting an object, including the shape and size of the required contact surface. This contact surface shape and size information can be provided to a 3D printing control system to create an adapter for the robot's armature. The results of AI-based shape recognition can identify objects as similar to types the platform has previously encountered. As an example, the object's parameters may be used by the platform to identify candidate objects in a library of objects that manage a fleet of robotic tasks. The library may further indicate that Sling has been used successfully in one or more previous encounters with objects of this class. In embodiments, control of a robotic system for 3D printing may be directed to create an appropriate sling to be used by one or more robots assigned to perform object-specific tasks to lift and transport an object. In another exemplary embodiment of a robot management platform with both 3D printing control and artificial intelligence-based shape recognition capabilities, repair of an object is accomplished by the use of vision and other sensors of the robotic system, such that the handle of the object to be repaired is broken and , which may be accomplished by determining that performance of the repairs as indicated thereby is impossible. Based on the determination of this unexpected condition, assign the current repair to instruct the robotic control system for 3D printing to create a replacement handle or perform repairs on the handle (e.g., repair a break in the structural part of the handle). A supplementary set of robot movements can be created for . These remedial actions may be determined based on an evaluation of the object to be repaired and incorporated into the current instance of the object repair process, even when the cause of the failure requiring the repair task is something other than the handle.

According to some example implementations, a robotic fleet management platform with a compliant (e.g., liquid) lens vision system may also focus on completing a set of tasks using quality of task completion as one of one or more training factors. You can have an AI loop-based training and learning system that can have. In embodiments, adaptive lens vision systems may be configured, controlled, and adapted through the use of artificial intelligence to improve image formation. Feedback from the AI loop-based training and learning system can be used as one element of feedback to adjust the adaptive lens for improved image formation. In embodiments, a combined AI system may facilitate adapting adaptive lenses to improve the quality of task completion. Factors such as destruction of task objects and/or robot components may suggest that image formation needs improvement, based on robot movements with a track record of success (e.g., not destroying objects). Robotic vision systems with adaptive lens technology can further improve robot behavior by using loop-based learning capabilities to train themselves to detect and provide guidance to avoid task execution hazards, such as objects along the path. there is.

According to some example implementations, the robotic fleet management platform with quantum optimization of thermal and energy factors in robotic systems may also include chip sensor systems (e.g., organ-on-a-chip, etc.) that provide biological detection and evaluation. You can have it. In some of these embodiments, the system detects radioactivity to evaluate conditions associated with the use and deployment of radioactive materials (e.g., as fuel for electric generators). The sensitivity of radiometric sensors and many other types of sensors can be affected by temperature conditions proximate to the sensing element. Ensuring that thermal factors are automatically and appropriately addressed throughout the robot's task assignment and over the life of the robot (or at least the sensing element) can improve sensitivity and thus reduce potentially hazardous levels of radiation with greater safety. This could potentially make detection easier. Maintaining thermal stability could provide additional benefits to other robotic sensing capabilities, such as chip-based medical diagnostic sensors, chip-based medical laboratory testing, etc.

According to some example implementations, a robotic fleet management platform with a computer vision infrastructure for tracking and managing general robotic assets may also have shared economy robotic resource scheduling and routing capabilities. In embodiments, a computer vision robot tracking infrastructure may provide context data to an autonomous robot resource routing embodiment of a shared economy robot resource scheduling and routing capability. In an example, the computer vision infrastructure may detect non-compliant robot behavior that may indicate a need for routing of robotic resources to replace/support/adjust one or more robots at the source of the detected non-compliant behavior. Additionally, a computer vision infrastructure for governing robotic assets provides evidence of task completion for autonomously routed robotic resources to facilitate the deployment of routed resources and automated billing of task completion by routed resources. can do. Such evidence may be provided to third parties (e.g. , other robot fleet platforms) can further substantiate claims.

143 illustrates an example data flow of an MPR 12100 adapted to harvest agricultural crops. In embodiments, data flow in MPR 12100 is partially driven by vision and sensing system 12112, motion planning system 12158, robot control system 12150, and module management system 12148 of MPR 12100. It is executed as In this example, MPR 12100 may be employed as part of a fleet of robots servicing agricultural environments (e.g., outdoor agricultural facilities, indoor agricultural facilities, containers configured to grow agricultural products, etc.). MPR (12100) selectively harvests agricultural units. In an embodiment, MPR 12100 may be configured to identify units within an agricultural farm that are ready to be harvested. In response, MPR 12100 navigates the environment to reach those units to perform the harvest task. MPR 12100 may utilize a vision and sensing system 12112 to identify which units are ready to be harvested. In response, motion planning system 12158 may generate a motion plan to navigate the environment and reach the location of units determined to be ready to be harvested. In embodiments, MPR 12100 may select one or more appropriate end effectors 12124 from modular system 12120 for selective harvesting of ready units.

At 12652, MPR 12100 may capture image data from an agricultural environment. Image data may include one or more images captured by cameras and/or other image sensors of vision and sensing system 12112. In some embodiments, one or more of the images may be captured using camera 12608 with a conformable varifocal liquid lens 12612. Images may be captured from multiple different viewpoints and may include one or more aerial images. In embodiments, cameras and/or other image sensors may be integrated within the housing of MPR 12100, such that MPR 12100 navigates the agricultural environment to capture images. Additionally or alternatively, one or more cameras and/or other image sensors may be integrated into other robots or located in various areas of the agricultural environment such that captured images are communicated to MPR 12100 for processing. do. Image data may be provided to machine vision system 12618 in intelligence layer 12140.

At 12654, machine vision system 12618 may then analyze the image data to identify obstacles within the environment. In embodiments, machine vision system 12618 may utilize one or more neural network models (e.g., CNN or RCNN) to detect various objects and obstacles in images. In embodiments, motion planning system 12158 builds a motion planning graph that represents the geometry of the environment as well as different possible paths that may lead to crops to be harvested (target objects).

At 12656, the motion planning system generates a motion plan for the MPR 12100 to identify the optimal path for it. In embodiments, motion planning system 12158 collaborates with intelligence layer 12140 to determine the optimal path after considering a cost function based on collision evaluation. The optimal path may be communicated to robot control system 12150. At 12658, robotic control system 12150 drives actuators within transport system 12110 to enable MPR 12100 to navigate to the location of units to be harvested. Navigation actions (e.g., forward, backward, right, rotation, etc.) for MPR 12100 are based on a trained navigation policy (machine learning algorithm). At 12660, the reinforcement learning system intelligence layer 12140 of 12622 may collect result data for the navigation policy to update and improve the policy.

Once MPR 12100 is close to the crops to be harvested, the controller can drive the movable arm to click additional images using a camera 12608 mounted on the end of the arm. The movable arm must be flexible in a dynamic environment and accurate enough not to damage the crop during movement.

At 12662, MPR 12100 may capture additional image data. For example, camera 12608 and/or other sensors mounted on the arm of MPR 12100 may capture images or other sensor data in an area proximate to the arm of MPR 12100. At 12664, machine vision system 12618 analyzes the images (using neural network models, including CNN or RCNN) to identify one or more crops to be harvested. At 12666, the images are analyzed by a motion planning system 12158 to determine the optimal path, as well as build a motion plan for the robot arm to rotate or move the arm without damaging the crops. . At 12668, the optimal path may be communicated to control system 12150, which may drive actuators within arm and end effector 12124 to harvest units of the crop. At 12670, the resulting data for the harvesting policy is collected by the reinforcement learning system 12622 in the intelligence layer 12140 as feedback to improve the harvesting policy.

SMART CONTRAINERS

In some embodiments, the value chain network may include a smart intermodal shipping container system 13000 that enables various capabilities noted above, including smart container fleet management services within the value chain network. System 1300 (as referenced above, and except where context indicates otherwise, any system, method, manufactured article, device, machine, equipment, algorithm, part, component, service, module, workflow, process, structure) , products, and other elements) are arranged to implement or enable a range of highly functional smart shipping containers, as well as forming a fleet of smart container operating units to, among other capabilities, e.g. within containers (autonomously, under remote control, or in combination) to coordinate operations with factories, ships, shipping docks, ports, warehouses, and transportation infrastructure (e.g. trucks, rail, etc.) To undertake operations (e.g., packaging, finishing, manufacturing of goods, moving or arranging items and/or handling storage conditions), and/or autonomously or remotely controlled mobility (or their It is configured to connect, combine and coordinate with other elements and entities of the value chain network, in order to undertake combinations. In embodiments, a smart container may physically store and transport cargo, where the cargo may include any merchandise, promotional items, ingredients, liquids, solids, powders, gases, food, etc. may refer to fields, and many other things.

In embodiments, intermodal smart containers 13026 may include containers of many different types, classes, sizes, weights, materials, shapes, capabilities, etc. In embodiments, smart containers may include standard rectangular containers, such as those with dimensions of 8-ft wide by 20-ft or 40-ft long, or non-standard containers. In embodiments, smart containers include 40-ft high-cube containers, 45-ft high-cube containers, 48-ft high-cube containers, 53-ft high-cube containers, etc. Additionally, in certain preferred embodiments, any module designed by supply chain and transportation operators for compatibility with infrastructure elements such as docks, factories, ports, trucks, trains, etc. Can contain containers of different sizes. In embodiments, smart containers may include tank containers (e.g., for liquids, gases, solids, powders, etc.), general purpose dry vans (e.g., boxes, crates, cases, sacks, pallets, pallets, drums, etc.). etc.), rolling floor containers, garmenttainers (e.g. for shipping garments on hangers), ventilated containers (passively or actively ventilated), temperature controlled containers (e.g. e.g. insulated, refrigerated and/or heated), bulk containers, open-top containers, open-side containers, log cradles, platform-based containers (e.g. , flat rack and bolster containers), rotating and/or mixing containers (e.g. cement mixers), air containers (unit load devices), automotive containers (e.g. transporting passenger vehicles) for), bioprotective containers (for shipping toxic or bioactive substances, such as those containing positive or negative pressure systems to regulate air flow), and many other containers. In some embodiments, smart containers may be smart packages (e.g., small 16-in x 12-in x 12-in boxes and many other smaller and larger sizes). In some embodiments, smart containers may be implemented, included, integrated with, or use robots configured to manipulate, transport, store, or deliver payload. In embodiments, a smart container may be configured to use external or internal walls, housing elements, or other interior elements, for example, to increase or decrease the volume of the container or to change the dimensions of one or more partitions of space within the container. May include mechanisms that allow elements to expand or contract. Smart containers may have self-assembly and/or self-disassembling mechanisms. In embodiments, smart containers are organic-like (e.g., to enable thermal management by providing a conductive or convective interface to a heating or cooling environment or an active heating or cooling element). ) or shapes that include linear, non-linear, and irregular shapes, such as biomimetic shapes, may be rectangular solids, cubes, spheres, cylinders, or other shapes.

Smart Container (13026) can be made of corrugated weathering steel, steel alloy, stainless steel, aluminum, cast iron, concrete, ceramic material(s), other alloys, glass, other metals, plastic, plywood, bamboo, cardboard ( It may consist of or include a diverse set of materials such as cardboard, wood, and/or many other materials. In embodiments, the smart container may be biodegradable. In embodiments, the smart container may be a 3D printed smart container or may include 3D printed elements. In embodiments, the 3D printed smart container may be printed as a single integrated unit. In embodiments, a 3D printed smart container may have embedded 3D printed electronics. In embodiments, a smart container includes a 3D printer or other additive manufacturing facility that prints one or more components, tools, accessories, etc. for integration into and/or use by the container.

In embodiments, intermodal smart container 13026 may be autonomous and/or self-driving. For example, intermodal smart containers can be configured to autonomously traverse terrestrial, underground, marine, subsea, airborne, and/or space environments. In embodiments, a smart container may include retractable or non-retractable wheels (e.g., for roads, terrain, rail, etc.) to support transportation in different environments; It may consist of continuous track systems, skis, rails, propellers, propulsion systems, bridges, etc. Additionally or alternatively, smart containers can be transported by traditional methods such as rail, truck, container ship, and air. For example, a smart container with retractable wheels could autonomously drive across a container terminal, drive up a ramp onto a container ship, and drive to a specific location on a container ship. and can be transported to other container terminals via container ships. Smart containers can also be configured to be transported by hyperloop systems and networks.

In some embodiments, smart containers 13026 may be configured to be self-stacking. For example, a smart container may have mechanical stacking rails on the sides that allow the container to slide up and down and/or across other containers. In embodiments, the rails may be electromagnetic rails. Additionally or alternatively, smart containers may be comprised of collapsible or non-collapsible mechanical lift systems, container handling devices, etc. For example, smart containers can be equipped with lift systems such as scissor lifts (e.g. hydraulic scissor lifts, diesel scissor lifts, electric scissor lifts, rough terrain lifts) that enable self-stacking of smart shipping containers. Scissor lift, or pneumatic scissor lift), foldable extendable legs, etc. In other examples, smart containers may consist of cranes, forklifts, reach stackers, etc.

In an embodiment, the fleet management system 13002 of the smart intermodal container system 13000 may receive an order for cargo storage and/or transportation services (e.g., from a client device) and store and/or transport services to be performed upon completion of the order. /or identifies the transportation service. In some embodiments, a user may be provided with a GUI on a client device to provide one or more freight service requirement parameter values. For example, the GUI can help users determine service timing requirements (e.g., how quickly the shipment needs to be delivered), origin of shipment, whether the shipment is received at a terminal/ramp or another location, Destination of the shipment, whether delivery is to a terminal/ramp or another location, type of container required, number of containers required, container usage requirements (e.g., full container (FCL) vs. shared container (LCL)), FCL Container size requirements for shipment (e.g., 20-ft container, 40-ft container, 40-ft high-cube, tank, etc.), cargo descriptions for LCL shipment (e.g., package number, total volume, total weight, etc.), whether the cargo contains personal effects (person effects), etc. In response to determining requirements for cargo storage and/or transportation service orders, the smart container system's fleet management system 13002 may determine a smart container fleet configuration, including a set of smart container operating units, and Can be assigned to cargo storage and/or transportation service orders. As used herein, a smart container operating unit may refer to an individual smart container, a team of smart containers, a fleet of smart containers that operate to complete a request, etc. As will be discussed, in some embodiments, smart container fleet management system 13002 may configure one or more smart containers to execute cargo storage and/or transportation service orders and/or operate in certain types of environments as part of a fleet configuration. Configuration can be defined. As discussed, a smart container can be composed of various modules that enable the smart container to perform specific tasks. For example, smart containers include specialized devices and systems such as IoT devices (including camera- and sensor-based devices), edge network devices, chipsets, chips, and data storage, computation, and data storage that enable smart containers to perform intelligent tasks. Other devices with processing and/or connectivity capabilities; Special sensors for specific environments; liquid lenses to enable certain machine vision functions; Special robots that perform specific tasks, special tools and/or systems that are specific tasks (e.g. lighting, irrigation systems, heating, cooling, 3D printers, clamps, grippers, drills, lifts, cranes, conveyors, mixers) , forklift, etc.); smart lamp; and/or other modules that configure the smart container to perform a specific task or set of tasks may be provisioned.

In some embodiments, the fleet management system 13002 of the smart intermodal container system may define a set of workflows, where the workflows define the order in which specific cargo storage and/or transportation services or tasks are performed and each You can define smart container operating unit(s) assigned to a service or task. In some embodiments, a smart container fleet management system may perform workflow simulations to iteratively redefine fleet configurations and/or workflows to substantially optimize the operation of a smart container fleet. For example, fleet composition and/or workflows can be iteratively adjusted to reduce costs, improve logistics efficiency, reduce overall delivery time, reduce carbon emissions, etc. Once the fleet configuration and workflow is complete, the smart container fleet management system can deploy the fleet. In some embodiments, a smart container fleet management system may facilitate logistics associated with the provision of smart container operating units and/or smart container components, and/or support of resources. Additionally, in some embodiments, smart container fleet management systems may utilize additive manufacturing capabilities, such as 3D printers or other capabilities described herein, to facilitate accommodating localized preferences with rapid customization, so that items can be It can be 3D printed inside a smart container as it approaches its destination. In embodiments, a smart container fleet management system includes the status of smart container operating units, performance of smart container units (e.g., timing performance, financial performance, etc.), status of cargo contained within smart containers, etc. can be monitored. In some of these embodiments, smart container system 13000 may automate maintenance of smart containers and/or resources to ensure efficient use of available inventory and/or reduce downtime.

In embodiments, smart container system 13000 uses a machine vision-based artificial intelligence system trained on a training set of data to recognize items by type, for example, by classifying items. thing); By determining appropriate storage conditions for the set (e.g., an artificial intelligence system, such as one trained on, inter alia, expert interactions with a model, expert commands or results from the setting and/or actions of storage conditions) may include considering the value of items that may have different optimal storage conditions, e.g., based on an economic or other model, which may also be performed by; and temperature profiles, humidity profiles, motion profiles that can be taken as a recommendation for further consideration (e.g., by a human operator or other entity) or as input to an autonomous or semi-autonomous control system for the environment. By providing a set of instructions for storage, such as a set of items, the storage conditions of a set of items can be automatically controlled. AI or expert systems may include factors related to the sensitivity of the contents to shaking, temperature changes, humidity, radiation, chemical factors (e.g., corrosion by salts), and many others; Factors related to the price of the contents (such as the price of the contents, costs, or profit margins, including based on market factors, contract terms, risk allocation, etc.); and transportation factors (e.g., equipment conditions, road, waterway, airway or rail conditions, weather conditions, and the like); and other factors. Over time, input factors can be added or removed depending on the success of artificial intelligence systems (including model-based, deep learning, or other systems), which are fed into expert models and expert-labeled trained based on data sets and/or trained based on expert interactions (eg, using robotic process automation) and/or trained based on results from usage. Accordingly, a self-regulating or autonomous storage conditions system, or a semi-autonomous storage conditions system, may be integrated with or integrated into a container system.

In some embodiments, the smart container system may monitor the status of the smart container operating unit and /Or it can support a digital twin that represents performance. The digital twin served by the smart intermodal container system can be adapted for a variety of purposes. In some example embodiments, a digital twin may be configured to provide the state of a smart container fleet, including individual smart containers within the fleet. In these examples, a user can drill down to individual smart containers within a team or fleet of smart containers to view the status of the smart containers. For example, the user can determine the smart container's battery life, availability of smart container energy sources and/or charging stations, location of the smart container, movement options for the smart container, status of cargo within the smart container, and task completion of the smart container. You can view the status, maintenance alerts of smart containers, etc. In some example embodiments, the smart container system may be configured to determine the location of available smart container infrastructure and/or transportation modes (e.g., container ships, submarines, spacecraft, hyperloops, rail, trucks, etc.), facilities (e.g. , container terminals, shipyards, storage areas, etc.), objects, other smart containers, sensor readings of the environment, etc., can serve an environmental digital twin that depicts the environment of a smart container fleet with real-time information. In these embodiments, a user can utilize the environmental digital twin to provide remote control commands to a smart container, a team of smart containers, or a fleet of smart containers. For example, a smart container or team of smart containers may encounter an unidentified object that disrupts the route and may need to make decisions regarding re-routing. In some embodiments, a smart container fleet management system may acquire relevant data (e.g., LIDAR data, video feeds, environmental maps, etc.) that can be depicted in an environmental digital twin. Users can view the current scenario in the environmental digital twin and provide commands to the smart container fleet on how to proceed given the scenario presented in the environmental digital twin. The foregoing are non-limiting examples of digital twins that can be used in connection with a smart container fleet management system, and other examples are discussed below.

144 illustrates an example environment of smart container system 13000. In an embodiment, smart container system 13000 includes a fleet management system 13002, a data processing system 13024, and an intelligence service 13004 (e.g., system level intelligence service 13004). In embodiments, fleet management system 13002 configures and manages smart container operating units 13040 and/or cargo storage and/or transportation services and/or tasks performed by smart container operating units 13040. As will be discussed, smart container operating unit 13040 may refer to an individual smart container, an individual smart container task assembly 13050, a smart container fleet 13060, and/or a smart container fleet support unit 13080.

In embodiments, fleet management system 13002 includes a communications management system 13010, a remote control system 13012, a resource provisioning system 13014, a logistics system 13016, a work organization system 13018, a fleet organization system ( 13020), including an order execution, monitoring, and reporting system 13022 (also referred to as “order execution system” 13022), a human interface system 13038, and a maintenance management system 13028. However, it is not limited to this. In an embodiment, communication management system 13010 is configured to facilitate fleet management system communication including elements external to smart container system 13000. In embodiments, fleet management system communications include satellite communications. In an embodiment, remote control system 13012 is configured to remotely manage and enable control of smart container operating units and fleet resources. In an embodiment, resource provisioning system 13014 is configured to handle allocation and access to fleet resources (e.g., smart container operating units). In an embodiment, logistics system 13016 coordinates the use and transportation of fleet resources and supplies to smart container operating units. In embodiments, maintenance management system 13028 facilitates coordinated and timely maintenance of fleet resources. In an embodiment, work organization system 13018 generates an order execution plan based on freight storage and/or transportation service orders. In embodiments, fleet configuration system 13020 configures smart container operating units (e.g., individual smart containers and/or smart container fleets) to complete order execution plans. In an embodiment, order execution system 13022 may be configured to execute an order execution plan and address shipping and fleet-related reporting requirements while ensuring efficient use of fleet resources (e.g., in accordance with the order execution plan) within a smart container operating unit. Performs, monitors and/or reports on cargo storage and/or transportation services performed by. In embodiments, a human interface system provides an interface through which a human user can interface with a smart container operating unit.

As mentioned, smart container operating unit 13040 may refer to an individual smart container 13026, an individual smart container task assembly 13050, a smart container fleet 13060, and/or a smart container fleet support unit 13080. there is.

As shown in FIG. 145 , smart container 13026 may include a baseline system 13106, a smart container control system 13104, and a smart container security system 13046. In an embodiment, smart container control system 13104 includes a data processing system 13024 and intelligence services 13004. As will be discussed, a data processing system may include data processing resources that may be centralized and/or distributed among a team or fleet of smart containers. Additionally or alternatively, the data processing resources may include general purpose chipsets, specialty chipsets, and/or configurable chipsets. As discussed, intelligence services 13004 perform intelligent related tasks on behalf of a smart container or collection of smart containers (e.g., a task assembly or fleet). For example, the intelligence service 13004 may perform tasks such as artificial intelligence, machine learning, natural language processing, machine vision, analytics, etc., in the performance of complex data structures (e.g., digital twins) and heterogeneous data. Data sources (e.g., from IoT, edge and other network-enabled devices, on-premises and cloud-deployed databases and other resources, and/or APIs, among many others, event streams, logs, or other data sources). Smart container-level and fleet-level intelligence services are discussed in more detail below. In embodiments, smart container security system 13046 performs security-related functions on behalf of a smart container or collection of smart containers (e.g., a task assembly or fleet). These security-related functions may include autonomous adaptive and non-adaptive security functions as well as passive security functions.

In embodiments, the baseline system 13106 of the smart container 13026 may include an energy storage and power distribution system 13112, an enclosure 13114, an electro-mechanical and/or electro-fluidic system 13116, It may include a transportation system 13118, a vision and sensing system 13120, and/or a rescue system 13122. As discussed further below, the configuration of the baseline system of smart container 13026 includes the types of cargo storage and/or transportation services and/or tasks that smart container 13026 is configured to perform. It depends on the types of environments in which the container is intended to operate. For example, a smart container configured to operate in deep-sea conditions may have a different baseline system than a smart container configured to operate in arctic conditions or an airborne smart container.

In embodiments, smart container 13026 may further include a modular system 13102 that allows the smart container to be composed of various hardware and/or software components. In this way, the smart container 13026 may be equipped with different accessories, sensor sets, depending on the scope of cargo storage and/or transportation services and/or tasks for which the smart container is configured and/or the environments in which the smart container is configured to operate. , chipsets, motivation adapters, etc. In embodiments, module system 13102 may include control module interfaces 13108 and physical module interfaces 13110. Control module interfaces 13108 and physical module interfaces 13110 may refer to mechanical, electrical and/or digital interfaces that receive auxiliary components to configure smart container 13026 to perform specific tasks. In embodiments, control module interfaces 13108 accommodate (or otherwise “connect”) auxiliary components that modify one or more features related to control of smart container 13026. These may include chipsets (e.g., AI chipsets, machine learning chipsets, machine vision chipsets, communication chipsets, etc.), sensor modules, communication modules, AI modules, security modules, computing modules, etc. In embodiments, physical module interface 13110 accommodates (or otherwise connects to) physical actions that may be taken by smart container 13026 and/or auxiliary physical modules that change the physical behavior of the smart container. Examples of physical modules may include, but are not limited to, wheels, robotic arms, sorting systems, packaging systems, 3D printers, cranes, lifts, power supplies, etc. As will be discussed, smart container 13026 may be reconfigured to perform one or more tasks upon completion of cargo storage and/or transportation service orders. In these embodiments, smart container system 13000 may define an order execution plan and support smart containers or smart container fleets, and allow smart containers 13026 to be reconfigured to perform one or more specified tasks in the order execution plan. , one or more modules may be provisioned to the smart container 13026.

Referring back to Figure 144, individual smart container task assembly 13050 may refer to a collection of one or more individual smart containers assigned to order cargo storage and/or transportation services. Smart containers Smart containers within a task assembly may include any combination of smart containers (eg, a 40-ft smart container and a tanker). In some embodiments, an individual smart container task assembly 13050 may include a local manager that controls or otherwise provides instructions for smart containers within the task assembly 13050. In these embodiments, the local manager may be a designated supervisor smart container, robot, or human operator. In embodiments, the smart container supervisor may act as an edge device on behalf of the task assembly 13050 such that the smart container supervisor communicates with the smart container system 13000 or another suitable device or system and/or performs tasks. Assembly 13050 may be assigned specific processing and/or communication capabilities that enable it to perform data processing operations on its behalf. In embodiments, a smart container fleet is a collection of individual smart containers and/or task assemblies that collectively perform a set of tasks in completing cargo storage and/or transportation service orders. Additionally, flits may be arranged as flits of task groups, regional flits, and/or flits among flits. In embodiments, smart container fleets may be supported by a smart container fleet support unit. In embodiments, examples of smart container fleet support include on-premises edge and IoT devices, local data stores (and corresponding data interfaces), forklifts, cranes, mechanical lifts, reach stackers, maintenance support, May include charging stations and devices, replacement parts, batteries, accessories, docking stations, spare parts, and/or technicians.

In embodiments, smart container system 13000 may include data processing system 13024. In an embodiment, data processing system 13024 includes data handling service 13032 and data processing service 13030. Data handling service 13032 is configured to store, retrieve, and otherwise manage data in smart container system 13000. In embodiments, data handling service 13032 accesses a set of data stores 13042 and/or libraries 13044, whereby data handling service 13032 operates smart container system 13000 and/or smart container operations. Write and read data from data store 13042 and/or library 13044 on behalf of other components of unit 13040. In embodiments, data processing services 13030 perform data processing operations on behalf of other components of smart container system 13000 and/or smart container operating unit 13040. For example, the data processing service 13030 may perform database operations (e.g., table join, search, etc.), data fusion operations, etc. In embodiments, a data processing system may include distributed resources, centralized resources, and/or “on-chip” resources.

As shown in FIG. 146, intelligent service 13004 is configured to provide intelligent services to smart intermodal container system 13000 and/or other intelligent service clients. In some embodiments, the intelligent services 13004 framework may be at least partially replicated in the smart intermodal container system 13000, smart containers, VCN control towers, various VCN entities, and/or other Intel. In these embodiments, smart intermodal container system 13000 may include some or all of the capabilities of intelligent service 13004, whereby intelligent service 13004 may perform certain functions performed by subsystems of intelligent clients. is adapted to Additionally or alternatively, in some embodiments, intelligent service 13004 may be implemented as a set of microservices, such that different intelligent service clients can access intelligent service 13004 via one or more APIs exposed to the intelligent clients. You can utilize it. In these embodiments, intelligent services 13004 may be configured to perform various types of intelligent services that can be adapted for different intelligent service clients. In either of these configurations, intelligent service client 13324 may provide an intelligent request to intelligent service 13004, whereby the request may be configured to perform a specific intelligent task (e.g., detection/identification, decision, recommendation, reporting). , commands, classification, prediction, optimization, control actions, configuration actions, automation, training actions, NLP requests, etc.). In response, intelligent service 13004 executes the requested intelligent task and returns a response to intelligent service client 13324. Additionally or alternatively, in some embodiments, intelligence services 13004 may be implemented using one or more specialized chips configured to provide AI-assisted microservices such as image processing, diagnostics, location and orientation, chemical analysis, data processing, etc. You can. For example, smart containers with AI chipsets can provide navigation commands, container security (e.g., biometric access), environmental monitoring, generate insights related to container and/or cargo weight, and provide insights related to container capacity. Generate insights related to the structural integrity of containers, generate insights related to cargo damage, predict regulatory issues (e.g. customs), detect illegal and/or dangerous cargo, and implement autonomous control. may be configured to implement an intelligent service 13004 to provide, etc.

It is further noted that in some scenarios, artificial intelligence modules 13404 themselves may also be intelligence service clients. For example, rule-based intelligence module 13428 may request an intelligence task from machine learning module 13412 or neural network module 13414, such as requesting classification of an object appearing in a video and/or motion of an object. In this example, rule-based intelligence module 13428 may be an intelligence service client 13324 that uses classifications to decide whether to take a particular action. In another example, machine vision module 13422 may request a digital twin of a specified environment from digital twin module 13420, so that machine learning module 13412 may include features for training a machine learning model trained for the specific environment. You can request specific data from the digital twin.

In embodiments, an intelligence task may require certain types of data to respond to the request. For example, a machine vision task classifies objects that appear in an image or set of images and identifies features within the set of images (e.g., locations of items, other containers (e.g., Bureau International des Containers (BIC) code) , CSC approval plates, ISO 6346 reporting marks, etc.) to determine the presence of codes or information, faces, symbols or commands, expressions, parameters of motion, changes of state, etc.), etc. Requires images (and potentially other data). In other examples, NLP tasks require audio of speech and/or text data (and potentially other data) to determine meaning or other elements of the speech and/or text. In another example, an AI-based control task (e.g., decisions about the movement of a smart container) may require environmental data (e.g., a map, coordinates of known obstacles, images, etc.) and/or motion plans may be requested. In a platform level example, an analytics-based reporting task may require data from multiple different databases to generate a report. Accordingly, in embodiments, tasks that may be performed by intelligent service 13004 may require or benefit from specific intelligent service inputs 13470. In some embodiments, intelligent service 13004 may be configured to receive and/or request specific data from intelligent service inputs 13470 to perform a respective intelligent task. Additionally or alternatively, requesting intelligence service client 13324 may provide specific data in the request. For example, intelligent service 13004 may expose one or more APIs to intelligent service clients, whereby requesting clients 13324 provide specific data in the request via the API. Examples of intelligent service inputs may include, but are not limited to, sensors providing sensor data, video streams, audio streams, databases, data feeds, human input, and/or other suitable data.

In one embodiment, intelligent service 13004 may include an intelligent service controller 13402 and an artificial intelligence module 13404. In embodiments, artificial intelligence service 13004 receives an intelligence request from intelligent service client 13324 and any necessary data to process the request from intelligent service client 13324. In response to the request and specific data, one or more involved artificial intelligence modules 13404 perform intelligence tasks and output an “intelligence response.” Examples of intelligence modules 13304 responses include making decisions (e.g., control commands, suggested actions, machine-generated text, etc.), making predictions (e.g., predicted meaning of text snippets, suggested actions, etc.) predicted outcomes associated with, predicted failure states, etc.), classification (e.g., classification of objects in images, classification of spoken utterances, classified failure states based on sensor data, etc.), and/or other suitable functions of an artificial intelligence system. Can include outputs.

In embodiments, artificial intelligence modules 13404 include machine learning module 13412, rule-based module 13428, analytics module 13418, RPA module 13416, digital twin module 13420, and machine vision module ( 13422), an NLP module 13424, and/or a neural network module 13414. It is understood that the foregoing are non-limiting examples of artificial intelligence modules, and that some of the modules may be included or utilized by other artificial intelligence modules. For example, NLP module 13424 and machine vision module 13422 may utilize different neural networks that are part of neural network module 13414 in performing their respective functions.

In embodiments, intelligence modules 13304 include and provide access to a machine learning module 13412 that may be integrated into or accessed by one or more intelligent service clients. In embodiments, machine learning module 13412 may provide intelligent services, such as training machine learning models, utilizing machine learning models, enhancing machine learning models, performing various clustering techniques, feature extraction, etc. May provide machine-based learning capabilities, features, functions and algorithms for use by client 13324. In one example, machine learning module 13412 can provide machine learning computing, data storage, and feedback infrastructure for the simulation system. Machine learning module 13412 may also operate in cooperation with other modules, such as rule-based module 13428, machine vision module 13422, RPA module 13416, etc.

The machine learning module 13412 is configured to provide analytics, simulation, decision making, optimization, and predictive analytics related to data processing, data analysis, simulation generation, simulation prioritization, and simulation analysis of one or more components or subsystems of the intelligent service client 13324. You can define one or more machine learning models to perform. In embodiments, machine learning models are algorithms and/or statistical models that perform specific tasks without using explicit instructions, relying instead on patterns and inferences. Machine learning models build one or more mathematical models based on training data to make predictions and/or decisions without being explicitly programmed to perform specific tasks. In example implementations, machine learning models may perform classification, prediction, regression, clustering, anomaly detection, recommendation generation, decision-making, optimization, and/or other tasks.

In embodiments, machine learning models may perform various types of classification based on input data. Classification is a predictive modeling problem in which class labels are predicted for given examples of input data. For example, a machine learning model can perform binary classification, multi-class classification, or multi-label classification. In an embodiment, a machine learning model may output “confidence scores” that indicate the respective confidence associated with the classification of the input into each class. In embodiments, confidence scores may be compared to one or more thresholds to render discrete category predictions. In an embodiment, only a certain number of classes (e.g., one) with the relatively largest confidence score may be selected to render a discrete category prediction.

In embodiments, a machine learning model may output a probabilistic classification. For example, a machine learning model can predict a probability distribution over a set of classes, given sample input. Therefore, rather than outputting only the most likely class to which the sample input should belong, the machine learning model can output, for each class, the probability that the sample input belongs to that class. In an embodiment, the probability distribution for all possible classes may sum to 1. In embodiments, a softmax function, or other type of function or hierarchy, may be used to convert a set of real values each associated with a possible class into a set of real values in the range (0, 1) that sums to 1. In embodiments, the probabilities provided by the probability distribution may be compared to one or more thresholds to render discrete category predictions. In embodiments, only a certain number of classes (e.g., one) with the relatively greatest prediction probability may be selected to render discrete category predictions.

In embodiments, a machine learning model may perform regression to provide output data in the form of continuous numeric values. By way of example, a machine learning model may perform linear regression, polynomial regression, or non-linear regression. As described, in an embodiment, the softmax function or other function or layer squashes the set of real values each associated with two or more possible classes into a set of real values in the range (0, 1) that sums to 1 ( It can be used to squash). For example, a machine learning model may perform linear regression, polynomial regression, or nonlinear regression. As an example, a machine learning model may perform simple regression or multiple regression. As described above, in some implementations, the softmax function or other functions or hierarchies convert a set of real values, each associated with two or more possible classes, into a set of real values in the range (0, 1) whose sum is 1. Can be used for quashing.

In embodiments, machine learning models may perform various types of clustering. For example, a machine learning model may identify one or more previously defined clusters to which the input data is most likely to correspond. In some implementations where the machine learning model performs clustering, the machine learning model may be trained using unsupervised learning techniques.

In embodiments, the machine learning model may perform anomaly detection or outlier detection. For example, a machine learning model may identify input data that does not conform to expected patterns or other characteristics (e.g., as previously observed from previous input data). For example, anomaly detection may be used for fraud detection or system failure detection.

In some implementations, a machine learning model may provide output data in the form of one or more recommendations. For example, machine learning models may be included in a recommendation system or engine. For example, given input data that describes previous outcomes (e.g., scores, rankings, or grades representing amounts of success or enjoyment) for a particular entity, a machine learning model can, based on the prior outcomes, produce the desired outcome. It may output suggestions or recommendations of one or more additional entities that are expected to have .

As described above, a machine learning model may be or include one or more of a variety of different types of machine learning models. Examples of these different types of machine learning models are provided below for illustration. One or more of the example models described below may be used (e.g., combined) to provide output data in response to input data. Additional models other than the example models provided below may also be used.

In some implementations, the machine learning model includes one or more classifier models, such as, for example, a linear classification model; secondary classification model; It may be or may include this. A machine learning model may include one or more regression models, such as, for example, a simple linear regression model; multiple linear regression model; Logistic regression model; stepwise regression model; Multivariate adaptive regression spline; local estimated scatterplot smoothing models; It may be or may include this.

In some examples, the machine learning model is one or more decision tree-based models, such as, for example, classification and/or regression trees; Chi-square automatic interaction detection decision tree; decision stumps; conditional decision tree; It may be or may include this.

A machine learning model may be or include one or more kernel machines. In some implementations, a machine learning model may be or include one or more support vector machines. The machine learning model may include one or more instance-based learning models, such as, for example, a learning vector quantization model; Self-organizing map model; Locally weighted learning model; It may be or may include this. In some implementations, the machine learning model includes one or more nearest neighbor models, such as, for example, a k-nearest neighbor classification model; k-nearest neighbor regression model; It may be or may include this. The machine learning model may be a Bayesian model, such as one or more Bayesian models, such as, for example, a naive Bayes model ( ); Gaussian Naive Bayes Model; multinomial naive Bayes model; averaged 1-dependency estimator; Bayesian network; Bayesian trust network; Hidden Markov Model; It may be or may include this.

The machine learning model may include one or more clustering models, such as, for example, a k-means clustering model; k-median clustering model; expectation maximization model; hierarchical clustering model; It may include etc.

In some implementations, the machine learning model can be modified using one or more dimensionality reduction techniques, such as, for example, principal component analysis; Kernel principal component analysis; Graph-based kernel principal component analysis; principal component regression; Partial least squares regression; Sammon mapping; multidimensional scaling; projection pursuit; linear discriminant analysis; mixed discriminant analysis; second-order discriminant analysis; generalized discriminant analysis; flexible discriminant analysis; autoencoding; etc. can be performed.

In some implementations, a machine learning model can be implemented using one or more reinforcement learning techniques, such as a Markov decision process; dynamic programming; Q function or Q-learning; value function approach; deep Q-network; differentiable neural computer; A3C(asynchronous advantage actor-critics); deterministic policy gradient; etc. can be performed or these can be applied.

In embodiments, artificial intelligence module 13404 may include and/or provide access to neural network module 13414. In an embodiment, neural network module 13414 is configured to train, deploy, and/or utilize an artificial neural network (or “neural networks”) on behalf of intelligent service client 13324. Note that in the description, the term machine learning model may include a neural network, and as such, neural network module 13414 may be part of machine learning module 13412. In embodiments, neural network module 13414 may be configured to train a neural network that may be used by smart container management system 13000 and other intelligent service clients. Non-limiting examples of different types of neural networks include convolutional neural networks (CNNs), deep convolutional neural networks (DCNs), feedforward neural networks (including deep feedforward neural networks), and recurrent neural networks (RNNs) (including gated RNNs). (but not limited to), long-term/short-term memory (LTSM) neural networks, etc., as well as serially, in parallel, in acyclic (e.g., directed graph-based) flows, and/or intermediate decision nodes, recursive loops, etc. may include any of the neural network types described throughout this disclosure and the documents incorporated by reference herein, including but not limited to hybrids or combinations of the above, such as deployed in more complex flows that may include A neural network of a given type can take input from a data source or another neural network and provide output that is contained within the input set of the other neural network until the flow is complete and the final output is provided. In embodiments, neural network module 13414 may be utilized by other artificial intelligence modules 13404, such as machine vision module 13422, NLP module 13424, rule-based module 13428, digital twin module 13420, etc. there is. Exemplary applications of neural network module 13414 are described throughout this disclosure.

Neural networks contain groups of connected nodes, which may also be referred to as neurons or perceptrons. Neural networks can be organized into one or more layers. A neural network containing multiple layers may be referred to as a “deep” network. A deep network may include an input layer, an output layer, and one or more hidden layers located between the input layer and the output layer. Nodes in a neural network can be connected or completely disconnected.

In embodiments, the neural network may be or include one or more feed forward neural networks. In a feedforward network, the connections between nodes do not form cycles. For example, each connection may connect a node from a previous layer to a node from a later layer.

In embodiments, the neural network may be or include one or more recurrent neural networks. In some cases, at least some of the nodes of a recurrent neural network may form a cycle. Recurrent neural networks can be particularly useful for processing input data that is sequential in nature. In particular, in some cases, a recurrent neural network may transfer or retain information from a previous portion of an input data sequence to a subsequent portion of the input data sequence through the use of recursive or directed recurrent node connections.

In some examples, sequential input data may include time series data (e.g., time or images captured at different times subject to sensor data). For example, a recurrent neural network can analyze sensor data versus time to detect or predict swipe direction, perform handwriting recognition, etc. Sequential input data includes words within sentences (e.g., for natural language processing, conversation detection or processing, etc.); notes in musical composition; sequential actions taken by the user (e.g., to detect or predict sequential application usage); sequential object states; It may include etc. In some example embodiments, the recurrent neural network may be a long short-term (LSTM) recurrent neural network; gated recurrent units; Bidirectional recurrent neural network; Continuous-time recurrent neural network; neural history compressor; echo state network; Elman networks; Jordan networks; recursive neural network; Hopfield networks; fully recurrent network; sequence-to-sequence configurations; Includes etc.

In some examples, the neural network may be or include one or more acyclic sequence-to-sequence models based on self-attention, such as a Transformer network. Details of an example transformer neural network can be found at http://papers.nips.cc/paper/7181-attention-is-all-you-need.pdf.

In embodiments, the neural network may be or include one or more convolutional neural networks. In some cases, a convolutional neural network may include one or more convolutional layers that perform convolutions on input data using learned filters. A filter may also be referred to as a kernel. Convolutional neural networks can be particularly useful for vision problems, such as when the input data includes images such as still images or video. However, convolutional neural networks can also be applied for natural language processing.

In embodiments, the neural network may be or include one or more generative networks, such as, for example, generative adversarial networks. Generative networks can be used to generate new data, such as new images or other content.

In an embodiment, the neural network may be or include autoencoders. In some cases, the goal of an autoencoder is to learn a representation (e.g., low-dimensional encoding) for a set of data, typically for the purpose of dimensionality reduction. For example, in some cases, an autoencoder may attempt to encode input data and then provide output data that reconstructs the input data from the encoding. Recently, the autoencoder concept has become more widely used to learn generative models of data. In some cases, autoencoders may involve additional losses beyond reconstructing the input data.

In embodiments, the neural network may be one or more other types of artificial neural networks, such as, for example, a deep Boltzmann machine; deep trust network; Stacked autoencoder; It may be or may include this. Any of the neural networks described in this application can be combined (e.g., stacked) to form more complex networks.

In one embodiment, a neural network may include an input layer, a hidden layer, and an output layer, each layer including a plurality of nodes or neurons that respond to a different combination of inputs from the previous layer. Connections between neurons have numerical weights that determine how much relative influence the input has on the output value of that node. The input layer may include a plurality of input nodes that can provide information or input data (eg, sensor data, image data, text data, audio data, etc.) from the external world to the neural network. The input data may be from different sources and may include library data (x1), simulation data (x2), user input data (x3), training data (x4) and results data (x5). An input node may pass information to the next layer, and no computation may be performed by the input node. The hidden layer may include multiple nodes. Nodes in the hidden layer can process information from the input layer and pass the information to the output layer based on the weights of the connections between the input layer and the hidden layer. The output layer processes information based on the weights of the connections between the hidden layer and the output layer, computes and transmits the information from the network to the outside world, such as recognizing specific objects or activities, or predicting conditions or actions. It may contain output nodes that are responsible for doing this.

In embodiments, a neural network may include two or more hidden layers and may be referred to as a deep neural network. The layers are arranged such that the first layer detects a set of primitive patterns in the input (e.g., image) data, the second layer detects patterns of patterns, and the third layer detects patterns of corresponding patterns. It is composed. In some embodiments, a node within neural network 8840 may have connections to all nodes in the immediately previous layer and the immediately next layer. Accordingly, the layer may be referred to as a fully-connected layer. In some embodiments, nodes within neural network 8840 may have connections only to some of the nodes in the immediately previous layer and the immediately next layer. Accordingly, the layer may be referred to as a sparsely-connected layer. Each neuron in a neural network consists of a weighted linear combination of its inputs, and the calculations for each neural network layer can be described as the multiplication of the input matrix and the weighting matrix. Next, a bias matrix is added to the resulting product matrix to take into account the threshold of each neuron at the next level. Additionally, an activation function is applied to each resulting value, and the resulting value is placed in the matrix for the next layer. Therefore, the output from node i in the neural network can be expressed as:

yi=f (Σxiwi+bi)

Here, f is the activation function, Σxiwi is the weighted sum of the input matrices, and bi is the bias matrix.

The activation function determines the level of activity or excitation generated in a node as a result of an input signal of a certain magnitude. The purpose of an activation function is to introduce non-linearity into the output of a neural network node, since most real-world functions are non-linear and it is desirable for neurons to be able to learn such non-linear representations. Several activation functions can be used in artificial neural networks. One exemplary activation function is the sigmoid function σ(x), which is a continuous sigmoid monotonically increasing function that asymptotically approaches a fixed value as the input approaches plus or minus infinity. The sigmoid function σ(x) takes a real-valued input and converts it to a value between 0 and 1:

σ(x)=1/(1+exp(-x)).

Another example activation function is the tanh function, which takes a real-valued input and converts it to a value in the range [-1, 1]:

tanh(x)=2σ(2x)- 1

A third example activation function is a rectified linear unit (ReLU) function. The ReLU function takes a real-valued input and thresholds it above 0 (i.e. replaces negative values with 0):

f(x)=max(0, x).

Activation functions are provided as examples, and in various embodiments, the neural network 8840 may include identity, binary step, logistic, soft step, tanh, arctan, softsign, rectified linear unit (ReLU), and Ricky. commutated linear unit, parametrically commutated linear unit, random Leaky commutated linear unit, exponential linear unit, sigmoidally commutated linear activation unit, adaptive piecewise linear, softplus, bent identity, softexponential, It will be clear that a variety of activation functions may be used, including but not limited to sinusoidal, sinc, Gaussian, softmax, maxout, and/or combinations of activation functions.

In embodiments, the input layer may take external inputs x1, x2, x3, x4, and x5, which may be numeric values depending on the input data set. It will be appreciated that a node can contain dozens, hundreds, thousands, or more inputs. As discussed above, no computation is performed on the input layer, so the outputs are x1, x2, x3, x4 and x5 respectively, which are fed to the hidden layer. The output of a node in the hidden layer may depend on the output from the input layer (x1, x2, x3, x4, and x5) and the weights associated with the connections between the input layer and the hidden layer (w1, w2, w3, w4, and w5). You can.

The output from the node in the hidden layer can also be computed in a similar way and then fed to the node in the output layer. Nodes in the output layer can perform similar calculations (using weights v1, v2, and v3 associated with the connection) as nodes in the hidden layer.

As mentioned, connections between nodes within a neural network have weights associated with them, which determine how much relative influence the input value has on the output value of the node in question. Before the network is trained, random values are chosen for each weight. Weights are adjusted during the training process, and this adjustment of weights to determine the best set of weights that maximizes the accuracy of the neural network is referred to as training. For every input in the training data set, the output of the artificial neural network can be observed and compared to the expected output, and the error between the expected and observed outputs can be propagated backwards to the previous layer. The weights can be adjusted accordingly based on the error. This process is repeated until the output error is below a predetermined threshold.

In embodiments, backpropagation (e.g., backward propagation of errors) is used in conjunction with optimization methods such as gradient descent to adjust weights and update neural network properties. Backpropagation can be a supervised training technique that learns from labeled training data and errors at nodes by changing the parameters of the neural network to reduce errors. For example, the results of forward propagation (e.g., output activation value(s)) determined using training input data are compared to corresponding known reference output data to calculate the loss function gradient. The gradient can then be used in an optimization method to determine new updated weights in an attempt to minimize the loss function. For example, to measure error, the mean square error is determined using the equation:

E=(target-output) ²

To determine the gradient for weight “w”, the partial derivative of the error with respect to the weight can be determined, where:

Gradient=∂E/∂w

Computation of the partial derivative of the error with respect to the weights can flow backwards through the node levels of the neural network. Then, a portion of the gradient (e.g., ratio, percentage, etc.) is subtracted from the weight to determine the updated weight. That part can be designated as the learning rate "a". Accordingly, an exemplary equation for determining the updated weights is given by the formula:

w new =w old -α∂E/∂w

Choose the learning rate so that it is not too small (e.g., a rate that is too small may lead to slow convergence to the desired weights) and not too large (e.g., a rate that is too large may cause the weights to not converge to the desired weights). It has to be.

After adjusting the weights, the network should perform better than before for the same input because the weights have now been adjusted to minimize errors.

As mentioned, neural networks may include convolutional neural networks (CNNs). CNN is a specialized neural network for processing data with a known grid-like topology, such as image data. Therefore, CNNs are commonly used in classification, object recognition, and computer vision applications, but they can also be used for other types of pattern recognition, such as conversation and language processing.

Convolutional neural networks learn highly nonlinear mappings by interconnecting layers of artificial neurons arranged in many different layers with activation functions that form layer dependencies. It includes one or more convolutional layers interspersed with one or more subsampling layers and non-linear layers, which are typically followed by one or more fully connected layers.

In embodiments, a CNN includes, in that order, an input layer with an input image to be classified by the CNN, one or more activation or non-linear layers (e.g., ReLU), and one or more convolutional layers interspersed with pooling or subsampling layers. a hidden layer, and an output layer that typically includes one or more fully connected layers. The input image can be represented by a matrix of pixels and can have multiple channels. For example, a color image may have red, green, and blue channels, respectively representing the red, green, and blue (RGB) components of the input image. Each channel can be represented by a 2-D matrix of pixels with pixel values ranging from 0 to 255. Meanwhile, a gray scale image can have only one channel. The next section describes the processing of a single image channel using CNN. It will be appreciated that multiple channels may be handled in a similar manner.

As shown, the input image can be processed by a hidden layer comprising a set of convolutional and activation layers, each followed by a pooling layer.

The convolutional layer of a convolutional neural network serves as a feature extractor that can learn the input image and decompose it into hierarchical features. A convolution layer can perform a convolution operation on an input image, where a filter (also called a kernel or feature detector) can slide over the input image with a specific step size (also called a stride). For every position (or step), an element-wise multiplication between the filter matrix and the nested matrix within the input image is computed and summed to obtain a final value that represents a single element of the output matrix, forming the feature map. . A feature map refers to image data representing various features of input image data, and may have smaller dimensions compared to the input image. The activation or non-linear layers use different non-linear trigger functions for signaling distinct identifications of likely features on each hidden layer. The nonlinear layer uses a variety of specific functions to implement nonlinear triggering, including rectified linear units (ReLU), hyperbolic tangent, and absolute and sigmoid functions of hyperbolic tangent. In one implementation, ReLU activation implements the function y=max(x, 0) and keeps the input and output sizes of the layer the same. The advantage of using ReLU is that convolutional neural networks are trained many times faster. ReLU is a discontinuous, non-saturating activation function that is linear with respect to the input if the input value is greater than 0 and is 0 otherwise.

In one example, the first convolution and activation layer convolves the input image using a number of filters followed by a non-linearity operation (e.g., ReLU) to generate a number of output matrices (or feature maps). can be performed. The number of filters used may be referred to as the depth of the convolutional layer. Therefore, the first convolution and activation layer has a depth of 3 and uses 3 filters to generate 3 feature maps. The feature map may then be passed to a first pooling layer that may subsample or downsample the feature map using a pooling function to generate an output matrix. The pooling function replaces the feature maps with summary statistics to reduce the spatial dimension of the extracted feature maps, thereby reducing the number of parameters and computations in the network. Therefore, the pooling layer reduces the dimensionality of the feature map while retaining the most important information. Pooling functions can also be used to introduce translation invariance into a neural network, such that small transformations on the input do not change the pooled output. Different pooling functions can be used in the pooling layer, including max pooling, average pooling, and 12-norm pooling.

The output matrix may then be processed by a second convolution and activation layer to perform convolution and non-linear activation operations (e.g., ReLU) as described above to generate a feature map. In embodiments, the second convolution and activation layer may have a depth of 5. The feature maps can then be passed to a pooling layer, where the feature maps can be subsampled or downsampled to produce an output matrix.

The output matrix generated by the pooling layer is then processed by one or more fully connected layers that form part of the output layer of the CNN. A fully connected layer has full connectivity with all feature maps in the output matrix of the pooling layer. In an embodiment, a fully connected layer may take the output matrix generated by the pooling layer as input in the form of a vector and perform high-level decisions to output a feature vector containing information of the structure within the input image. In an embodiment, a fully connected layer may use a softmax function to classify objects in an input image into one of several categories. The softmax function can be used as an activation function in the output layer, taking a vector of real-valued scores and mapping it to a vector of values between 0 and 1, which sum to 1. In embodiments, other classifiers may be used, such as a support vector machine (SVM) classifier.

In embodiments, one or more normalization layers may be added to the CNN to normalize the output of the convolutional filter. A regularization layer can provide whitening or lateral suppression, avoid vanishing or exploding gradients, stabilize training, and enable learning at higher rates and faster convergence. In an embodiment, a normalization layer is added after the convolution layer, but before the activation layer.

Therefore, a CNN can be viewed as a multiple set of convolutional, activation, pooling, regularization and fully connected layers stacked together to learn, enhance and extract implicit features and patterns in input images. As used in this application, a layer operates with similar functionality by mathematical or other functional means to process the received input to generate/derive an output for the next layer to one or more other components for further processing within the CNN. Can refer to one or more components.

An initial layer (e.g., a convolutional layer) of the CNN may extract low-level features, such as edges and/or gradients, from the input image 8862. Subsequent layers can extract or detect progressively more complex features and patterns, such as the presence of curvature and texture in the image data, etc. The output of each layer can serve as the input to subsequent layers in a CNN to learn hierarchical feature representations from data in the input image. This allows convolutional neural networks to efficiently learn increasingly complex and abstract visual concepts.

Although only two convolutional layers are shown in the example, the present disclosure is not limited to the example architecture, and the CNN architecture may include any number of layers in total, and any number of layers for convolution, activation, and pooling. It can be included. For example, there have been many variations and improvements over the basic CNN model described above. Some examples are Alexnet, GoogLeNet, VGGNet (stacks many layers including a narrow convolutional layer followed by a max-pooling layer), residual network, or ResNet (which uses residual blocks and skip connections to learn residual mappings). used), DenseNet (connects each layer of the CNN to every other layer in a feed-forward fashion), Squeeze and excitation network (includes global context in features), and AmobeaNet (which uses evolutionary algorithms to search and find the optimal architecture for image recognition).

The training process of a convolutional neural network, such as a CNN, may be similar to the training process discussed in Figure 148 with respect to neural networks.

In an embodiment, all parameters and weights (including weights in filters and weights for fully connected layers) are initially assigned (e.g., randomly assigned). Then, during training, the training image or images in which objects are detected and classified are provided as input to the CNN, which performs a forward propagation step. That is, the CNN applies convolution, nonlinear activation, and pooling layers to each training image to determine a classification vector (i.e., detects and classifies each training image). These classification vectors are compared to predetermined classification vectors. The error (e.g., sum of squares of differences, log loss, softmax log loss) between the CNN's classification vector and a predetermined classification vector is determined. These errors are then used to update the CNN's weights and parameters in a backpropagation process, which may use gradient descent and may include one or more iterations. The training process is repeated for each training image in the training set.

The training process and inference process described above may be performed on hardware, software, or a combination of hardware and software. However, training a convolutional neural network, such as a CNN, or using a trained CNN for inference typically requires a significant amount of computational power, for example, to perform matrix multiplication or convolution. Therefore, specialized hardware circuits such as graphics processing units (GPUs), tensor processing units (TPUs), neural network processing units (NPUs), FPGAs, ASICs, or other highly parallel processing circuits may be used for training and/or inference. there is. Training and inference can be performed in the cloud, in a data center, or on the device.

In embodiments, an object detection model extends the functionality of a CNN-based image classification neural network model by not only classifying an object but also determining its location in the image in terms of its bounding box. Region-based CNN (R-CNN) method is used to extract regions of interest (ROI), where each ROI is a rectangle that can represent the boundaries of objects within the image. Conceptually, R-CNN operates in two steps. In the first step, the region proposal method generates all potential bounding box candidates in the image. In the second stage, for all proposals, a CNN classifier is applied to distinguish objects. Alternatively, a fast R-CNN architecture can be used that integrates feature extractors and classifiers into a unified network. Other faster R-CNNs can be used, which integrate Region Proposal Network (RPN) and fast R-CNN into an end-to-end trainable framework. Mask R-CNN adds instance segmentation, while Mesh R-CNN adds the ability to generate 3D meshes from 2D images.

In embodiments, artificial intelligence module 13404 may provide access to and/or integrate with robotic process automation (RPA) module 13416. RPA module 13416 may facilitate computer automation, among other things, creating and validating workflows. In embodiments, RPA module 13416 may monitor human interactions with various systems to learn patterns and processes performed by humans in the performance of individual tasks. This may include observation of human actions involving interactions with hardware elements, software interfaces, and other elements. Observations include, among many other examples, field observations of humans performing actual tasks, as well as simulations or other activities in which humans perform actions with the explicit intent of providing a training data set or input for the RPA system. It may involve observations, for example, where a human tags or labels a training data set with features that assist the RPA system in learning to recognize or classify features or objects. In embodiments, RPA module 13416 may learn to perform specific tasks based on learned patterns and processes, such that tasks may be performed by RPA module 13416 instead of or in support of a human decision maker. . Examples of RPA modules 13416 may include those in this disclosure and documents incorporated by reference herein, and may involve automation of any of the broad value chain network activities or entities described therein.

In embodiments, artificial intelligence module 13404 may include and/or provide access to analytics module 13418. In embodiments, analysis module 13418 is configured to perform various analysis processes on data output from value chain entities or other data sources. In example embodiments, analytics information generated by analytics module 13418 may facilitate quantification of system performance compared to a set of goals and/or metrics. Goals and/or metrics may be pre-configured, dynamically determined from operational results, etc. Examples of analysis processes that may be performed by analysis module 13418 are discussed below and in documents incorporated by reference into this application. In some example implementations, the analysis process involves coordination of value chain activities and demand intelligence, such as (among many others) involving anticipating demand for a set of related items by location and time. This may include tracking certain goals and/or specific metrics.

In embodiments, artificial intelligence module 13404 may include and/or provide access to digital twin module 13420. Digital twin module 13420 may include any of the wide range of features and capabilities described in this application. In an embodiment, the digital twin module 13420 may, among other things, support a smart container digital twin 13504, a digital twin of the physical shipping environment (shipping yard, container port, etc.), a transportation mode (container ship, truck, rail, etc.) ), digital twin of a smart container operating unit, logistics digital twin, organizational digital twin, role-based digital twin, etc. In embodiments, digital twin module 13420 may be configured according to digital twin systems and/or modules described elsewhere throughout this disclosure. In an example embodiment, digital twin module 13420 may be configured to create a digital twin requested by an intelligent service client. Additionally, the digital twin module 13420 may be configured with an interface such as an API for receiving information from an external data source. For example, digital twin module 13420 may receive real-time data from sensor systems in smart containers, machines, vehicles, robots, or other devices, and/or sensor systems in the physical environment in which the devices operate. In embodiments, digital twin module 13420 may receive digital twin data from other suitable data sources, such as third party services (e.g., weather services, traffic data services, logistics systems and databases, etc.). In an embodiment, the digital twin module 13420 may support value chain network entities such as supply chain infrastructure entities, transportation or logistics entities, containers, goods, etc., as well as customers, merchants, stores, points-of-sale, etc. ), and may include digital twin data representing characteristics, status, etc. of demand entities such as points-of-use, etc. Digital twin module 13420 is or is integrated with an interface (e.g., a control tower or dashboard) for coordination of supply and demand, including coordination of automation within supply chain activities and demand management activities. , may be linked to it, or interact with it in some other way.

In embodiments, digital twin module 13420 may provide access to and manage a library of digital twins. The artificial intelligence module8 can access libraries to perform functions such as simulation of actions in a given environment in response to specific stimuli.

In embodiments, artificial intelligence module 13404 may include and/or provide access to machine vision module 13422. In an embodiment, machine vision module 13422 is configured to process images (e.g., captured by a camera, liquid lens system, etc.) to detect and classify objects within the image. In an embodiment, machine vision module 13422 receives one or more images (which may be single still shot images or frames of a video feed) and detects a “blob” in the image (e.g., using edge detection techniques, etc.). )". Machine vision module 13422 may then classify the blob. In some embodiments, machine vision module 13422 utilizes one or more machine learned image classification models and/or neural networks (e.g., convolutional neural networks) to classify blobs within an image. In some embodiments, machine vision module 13422 may perform feature extraction on the image and/or each blob within the image prior to classification. In some embodiments, machine vision module 13422 may utilize classifications made from previous images to confirm or update the classification(s) from previous images. For example, if an object detected in a previous frame was classified with a lower confidence score (e.g., the object is partially occluded or out of focus), machine vision module 13422 may If the object's classification can be determined with this higher confidence, the classification can be confirmed or updated. In an embodiment, machine vision module 13422 is configured to detect occlusion, such as an object that may be occluded by another object. In embodiments, machine vision module 13422 may provide additional input to assist with the image classification task, such as from radar, sonar, a digital twin of the environment (which can show the location of known objects), and/or the like. receives. In some embodiments, machine vision module 13422 may include or interface with a liquid lens. In these embodiments, liquid lenses facilitate improved machine vision (e.g., when focusing at multiple distances is required by the robot's environment and tasks) and/or other machine vision tasks enabled by the liquid lenses. You can do it.

In embodiments, artificial intelligence module 13404 may include and/or provide access to natural language processing (NLP) module 13424. In an embodiment, NLP module 13424 performs natural language tasks on behalf of intelligent service client 13436. Examples of natural language processing technologies include conversation recognition, conversation segmentation, speaker diarization, text-to-speech, lemmatization, morphological segmentation, and parts-of-speech tagging. , stemming, syntactic analysis, lexical analysis, etc., but is not limited to these. In embodiments, NLP module 13424 may enable voice commands received from a human. In embodiments, NLP module 13424 may receive an audio stream (e.g., from a microphone) and perform speech-to-text conversion on the audio stream to obtain a transcription of the audio stream. NLP module 13424 may process text (e.g., a transcription of an audio stream) to determine the meaning of the text using various NLP techniques (e.g., NLP models, neural networks, etc.). In embodiments, NLP module 13424 may determine an action or command spoken in the audio stream based on the results of the NLP. In an embodiment, the NLP module 13424 may output the results of NLP to the intelligent service client 13324.

In an embodiment, NLP module 13424 provides intelligent service client 13324 with the ability to communicate with a human user as well as parse one or more conversational voice commands provided by a human user to perform one or more tasks. The NLP module 13424 can perform conversation recognition to recognize voice commands, natural language understanding to parse and derive meaning from commands, and natural language generation to generate voice responses to the user when processing user commands. . In some embodiments, NLP module 13424 enables intelligent service client 13324 to understand commands and, upon successful completion of a task by intelligent service client 13324, provide a response to the user. In embodiments, NLP module 13424 may formulate and ask questions to the user when the context of the user's request is not completely clear. In embodiments, NLP module 13424 may use input received from one or more sensors, including vision sensors, location-based data (e.g., GPS data), to determine context information associated with the processed conversation or text data. Available.

In an embodiment, the NLP module 13424 uses neural networks such as recurrent neural networks, long short-term memory (LSTM), gated recurrent units (GRU), transformer neural networks, convolutional neural networks, etc. when performing NLP tasks.

In an example neural network for implementing NLP module 13424, the neural network is a transformer neural network. In this example, the transformer neural network includes three input stages and five output stages to transform the input sequence into an output sequence. Exemplary transformers include encoders and decoders. The encoder processes the input and the decoder generates output probabilities, for example. The encoder contains 3 stages and the decoder contains 5 stages. Encoder stage 1 represents the input as a sequence of positional encodings added to the embedded input. Encoder stage 2 and stage 3 include N layers (eg, N=6, etc.), where each layer includes a location-specific feedforward neural network (FNN) and an attention-based sublayer. Each attention-based sublayer in Encoder Stage 2 includes four linear projections and multi-head attention logic to be added and normalized to be fed to the location-specific FNN in Encoder Stage 3. Encoder stages 2 and 3 use residual concatenation followed by a normalization layer at the output.

An example decoder processes the output embedding as its input to help ensure that the prediction for position i depends on positions before/below i, with the output embedding shifted right by one position. In stage 2 of the decoder, the masked multi-head attention is modified to prevent positions from attending to subsequent positions. Stages 3-4 of the decoder include N layers (e.g., N=6, etc.), where each layer includes a location-specific FNN and two attention-based sublayers. Each attention-based sublayer in decoder stage 3 includes four linear projections and multi-head attention logic to be added and normalized to be fed to the location-specific FNN in decoder stage 4. Decoder stages 2-4 use residual concatenation followed by a normalization layer at the output. Decoder stage 5 provides a linear transformation followed by a softmax function to normalize the resulting vector of K numbers into a probability distribution containing K probabilities proportional to the exponents of the K input numbers.

In embodiments, artificial intelligence module 13404 may also include and/or provide access to a rules-based module 13428 that may be integrated into or accessed by intelligent service client 13324. In some embodiments, rule-based module 13428 may consist of programmatic logic that defines sets of rules and other conditions that trigger specific actions that may be performed in connection with an intelligent client. In embodiments, rule-based module 13428 may consist of program logic that receives input and determines whether one or more rules are satisfied based on the input. If the condition is met, the rule-based module 13428 determines an action to perform, which can be output to the request intelligence service client 13324. Data received by the rule-based engine may be received from intelligence service input 13470 and/or other inputs within artificial intelligence module 13404, such as machine vision module 13422, neural network module 13414, ML module 13412, etc. Can be requested from a module. For example, rule-based module 13428 may receive and respond to classification of objects in the smart container's field of view from a machine vision system and/or sensor data from the smart container's LiDAR sensor. Thus, the smart container can decide whether it should continue its path, change its course, or stop. In embodiments, rule-based module 13428 may be configured to make other appropriate rule-based decisions on behalf of individual intelligent service clients 13324, examples of which are discussed throughout this disclosure. In some embodiments, the rules-based engine may apply governance standards and/or analysis modules, which are described in more detail below.

In an embodiment, the artificial intelligence module 13404 interfaces with, and in response to, an intelligent service controller 13402 that is configured to determine the type of request issued by the intelligent service client 13324, such that the artificial intelligence module 13404 uses artificial intelligence when responding to the request. Module 13404 may determine the set of governance standards and/or analyzes to be applied. In an embodiment, intelligent service controller 13402 may include an analytics management module 13406, a set of analytics modules 13408, and a governance library 13410.

In an embodiment, the intelligent service controller 13402 is configured to determine the type of request issued by the intelligent service client 13324 and, in response, the governance standards to be applied by the artificial intelligence module 13404 when responding to the request. and/or determine a set of analyses. In an embodiment, intelligent service controller 13402 may include an analytics management module 13406, a set of analytics modules 13408, and a governance library 13410. In embodiments, analytics management module 13406 receives artificial intelligence module 13404 requests and determines governance standards and/or analytics associated with the requests. In embodiments, analysis management module 13406 may determine governance standards to apply to a request based on the type of decision requested and/or whether a particular analysis should be performed for the requested decision. For example, a request for a control decision that causes the intelligent service client 13324 to perform an action may be tied to a specific set of applicable governance standards, such as safety standards, legal standards, quality standards, etc., and/or One or more analyzes may be coupled to control decisions, such as analysis, safety analysis, engineering analysis, etc.

In some embodiments, analytics management module 13406 may determine a governance standard to apply to a decision request based on one or more conditions. Non-limiting examples of such conditions may include the type of decision being requested, the geographic location in which the decision is being made, the environment the decision will affect, current or predicted environmental conditions of the environment, etc. In embodiments, governance standards may be defined as a set of standards libraries stored in governance library 13410. In embodiments, a standard library may define conditions, thresholds, rules, recommendations, or other suitable parameters under which decisions may be analyzed. Examples of standard libraries may include legal standard libraries, regulatory standard libraries, quality standard libraries, engineering standard libraries, safety standard libraries, financial standard libraries, and/or other suitable types of standard libraries. In embodiments, governance library 13410 may include an index that indexes specific standards defined in individual standard libraries based on different conditions. Examples of conditions may be jurisdictions or geographic areas to which a particular standard applies, environmental conditions to which a particular standard applies, device types to which a particular standard applies, materials or products to which a particular standard applies, etc.

In some embodiments, the analytics management module 13406 may determine an appropriate set of standards that should be applied in connection with a particular decision, and the artificial intelligence module 13404 may utilize the associated governance standards in determining the decision. Module 13404 may be provided with an appropriate set of standards. In such embodiments, artificial intelligence module 13404 may be configured to apply standards in the decision-making process such that decisions output by artificial intelligence module 13404 are consistent with associated governance standards. It is understood that the standard library within the governance library may be defined by the platform provider, customer, and/or third party. Standards may be government standards, industry standards, customer standards, or other suitable sources. In embodiments, each set of standards may include a set of conditions that associate the respective set of standards such that the conditions can be used to determine which standard to apply in a given situation.

In some embodiments, analysis management module 13406 may determine one or more analyzes to be performed for a particular decision and may provide corresponding analysis modules 13408 to artificial intelligence module 13404 to perform these analyses, , the artificial intelligence module 13404 utilizes the corresponding analysis module 13408 to analyze the decision before outputting the decision to the requesting client. In embodiments, analysis module 13408 may include modules configured to perform specific analyzes on specific types of decisions, where each module is configured to perform a specific analysis on a specific type of decision by a processing system hosting an instance of intelligence service 13400. It runs. Non-limiting examples of analysis modules 13408 include risk analysis module(s), security analysis module(s), decision tree analysis module(s), ethics analysis module(s), failure mode and effects (FMEA) analysis module ( s), risk analysis module(s), quality analysis module(s), safety analysis module(s), regulatory analysis module(s), legal analysis module(s), and/or other suitable analysis module(s). .

In some embodiments, analysis management module 13406 is configured to determine which type of analysis to perform based on the type of decision requested by intelligent service client 13324. In some of these embodiments, analysis management module 13406 may include an index or other suitable mechanism to identify a set of analysis modules 13408 based on the requested decision type. In this embodiment, analysis management module 13406 can receive a decision type and determine a set of analysis modules 13408 to execute based on the decision type. Additionally or alternatively, one or more governance standards may define when certain analyzes are performed. For example, engineering standards may define which scenarios require FMEA analysis. In this example, an engineering standard may be tied to a request for a specific type of decision, and the engineering standard may define a scenario in which the FMEA analysis is performed. In this example, artificial intelligence module 13404 may execute a safety analysis module and/or a risk analysis module and determine alternative decisions about whether an action would violate a legal or safety standard. In response to analyzing the proposed decision, artificial intelligence module 13404 may optionally output proposed conditions based on the results of the performed analysis. If the decision is accepted, the artificial intelligence module 13404 may output the decision to the requesting intelligence service client 13324. If the proposed configuration is flagged by one or more of the analyzes, artificial intelligence module 13404 may determine an alternative decision and run the analysis on the alternative proposed decision until a matching decision is obtained.

Note here that, in some embodiments, one or more analysis modules 13408 may themselves be defined as standards, and one or more associated standards used together may include specific analyzes. For example, applicable safety standards may require a hazard analysis or one or more acceptable methods to be used. In this example, ISO standards for overall processes and documentation, and ASTM standards for narrowly defined procedures may be used to complete the risk analysis required by the safety governance standard.

As mentioned, the above-described framework of intelligent services 13004 may be applied and/or utilized by various entities in the value chain. For example, in some embodiments, a platform level intelligence system may be configured with the full capabilities of intelligence services 13400, and specific configurations of intelligence services 13400 may be provisioned for each value chain entity. Additionally, in some embodiments, an intelligent service client 13324, such as a smart container system 13000, may transfer an intelligent system task to a higher level value chain entity (e.g., when the intelligent service client 13324 is unable to perform the task autonomously). For example, it can be configured to escalate to edge-level or platform-level. Note that in some embodiments, intelligent service controller 13402 may direct intelligent tasks to lower-level components. Additionally, in some implementations, intelligence service 13004 may be configured to output a default action when a decision cannot be reached by intelligence service 13004 and/or a higher or lower level intelligence system. In some of these implementations, default decisions may be defined in rules and/or standard libraries.

In embodiments, a “set” of machine learning models may include a set with multiple members. In embodiments, a “set” of machine learning models may include hybrids of different types of models (e.g., hybrids of RNN and CNN).

In one example, a set of machine learning models can be used for smart container predictive maintenance. In this example, intelligence service 13004 may receive and receive order data, past order data, maintenance data, weather data, and/or video feeds from sensors within the smart container for user device 13094. A set of feature vectors can be created based on the generated data. Intelligence service 13004 may be configured to predict (e.g., using a combination of simulated and real-world data) when a particular smart container will require maintenance, e.g., based on a training data set of results. Feature vectors can be input to trained machine learning models. In embodiments, intelligence service 13004 may include an input set of training data representing the probability of required maintenance or predictions by a group of human experts and/or other systems or models.

In another example, a set of machine learning models can be used to predict traffic at a container terminal at a given point in time. In this example, intelligence service 13004 may receive historical container terminal traffic data, maritime data, news data, and weather data, and may generate feature vectors based on the received data. In embodiments, feature vectors may include other data, such as data characterizing container terminal layout elements on which traffic may depend. Intelligence service 13004 may input the feature vectors into a machine learning model trained (e.g., using combined simulated data and real-world data) to predict traffic at the container terminal.

In another example, a set of machine learning models may be used to detect counterfeit and/or illegal items being delivered. In this example, intelligence service 13004 may receive order data, shipper data, historical cargo data, and/or video feeds and other sensor data from sensors deployed inside the smart container, and the received data A feature vector can be generated based on . Intelligence service 13004 may input feature vectors into trained machine learning models (e.g., using a combination of simulated data and real-world data) to detect illegal and/or illegal items. In embodiments, detection of illegal and/or illegal items may be influenced by laws or regulations (e.g., training in awareness of legality), cultural factors (e.g., cultural norms that determine whether an item is considered illegal) or It may involve a set of separate models, each trained based on training data sets and/or feature vector inputs specific to the jurisdictional factors, including cases that change based on, etc.). In embodiments, training may be provided to a shipper or other user, for example, through a human expert, to describe illegal or illegal items, such as code words, euphemisms, etc. This may include providing information on alternative terms that may be employed (when describing the cargo to be delivered), etc. In embodiments, the model may use a word cloud, cluster of words, or other features to facilitate recognition of illegal or illegal items and/or words, images, or other elements used to characterize them. can be trained to provide them. As one non-limiting example, a SOM (self) is used to create mappings of entities, such as mapping entities, classes, objects, workflows, etc. -organizing map) can be used. Additionally or alternatively, a machine learning model may be configured to identify container contents.

In another example, a set of machine learning models can be used to provide decision support related to pricing of one or more cargo storage and/or transportation services (e.g., services requiring the use of smart containers). there is. For example, intelligence service 13004 may receive data from various sources described throughout this document and documents incorporated herein by reference, and generate a set of feature vectors based on the data received. there is. Intelligent service 13004 may be a trained machine (e.g., using a combination of simulated and real-world data) to provide decision support regarding the price of one or more services, such as based on a training data set of results. You can input feature vectors into the learning model. In embodiments, intelligent service 13004 may include an input set of training data representing decision support related to service pricing by a group of human experts and/or other systems or models. The data sources used to generate the set of feature vectors are order data, demand data, supply data, cost data, volatility data, price pattern data, order size data, order quantity data, geographic transaction data, maritime data, truck fleet data, This may include, but is not limited to, rail data, traffic data, weather data, social media sites, external data (e.g., news involving smart containers or shipping, etc.), and many others.

In another example, a set of machine learning models may be used to provide decision support related to loading and/or unloading of cargo. Intelligence service 13004 may receive order data (optionally including weight data, volumetric data, cargo description data, destination location, etc.) and/or video feeds from sensors placed outside and/or within the smart container. and a set of feature vectors can be generated based on the received data. Intelligence service 13004 may input a set of feature vectors into a trained machine learning model (e.g., using a combination of simulated and real-world data) to provide decision support related to cargo loading and/or unloading. there is. For example, machine learning models can be configured to provide decision support regarding the order in which specific cargo is loaded and/or unloaded, clustering of cargo, organization of cargo within a smart container, etc. In embodiments, a model or set of models may be trained by an expert in the loading and/or unloading of cargo.

In another example, a set of machine learning models can be used to determine regulatory compliance of a shipment. For example, intelligence service 13004 may receive data from various sources described throughout this document and documents incorporated herein by reference, and generate a set of feature vectors based on the data received. there is. Intelligence service 13004 may input the feature vector into a machine learning model trained to determine regulatory compliance. As a non-limiting example, regulatory compliance may include compliance with regulations that require documentation confirming that duties have been paid. In these example embodiments, the machine learning model may parse documents, commercial invoices, etc., e.g., to find verification of payment of requested duties.

In another example, a set of machine learning models can be used to categorize or classify cargo. For example, intelligence service 13004 may receive data from various sources described throughout this document and documents incorporated herein by reference, and generate a set of feature vectors based on the data received. there is. Intelligence service 13004 may, for example, convert feature vectors to a trained machine learning model (e.g., using a combination of simulated and real-world data) to categorize shipments based on a training data set of results. You can enter a set of . In embodiments, intelligence service 13004 may include an input set of training data representing categorizations or classifications of cargo by a set of human experts and/or by other systems or models. . Data sources and feature vectors used to categorize or classify cargo may include many types of transportation data, shipper profile data, and/or shipper profile data described herein, as well as external data sources that can assist in categorizing or classifying cargo. It can be included. In this example and other examples described herein, such artificial intelligence systems used for classification include recurrent neural networks (including gated recurrent neural networks), convolutional neural networks, combinations of recurrent neural networks and convolutional neural networks, or other types of neural networks or combinations or hybrids of types described herein or in documents incorporated herein by reference.

In another example, a set of machine learning models can be used to optimize smart container design. For example, intelligence service 13004 may receive data from various sources described throughout this document and documents incorporated herein by reference, and generate a set of feature vectors based on the data received. there is. Intelligence service 13004 may generate trained machine learning models (e.g., using a combination of simulated and real-world data) to optimize the design of a smart container, e.g., based on a training data set of results. You can input feature vectors. In embodiments, intelligence service 13004 may include an input set of training data representing smart container optimization by a fleet of human experts and/or other systems or models. Data sources and feature vectors used for optimization of marketplace efficiency may include many types of shipping data described herein that can assist in smart container design optimization. Such artificial intelligence systems used for optimization in this example and other examples described herein may include recurrent neural networks (including gated recurrent neural networks), convolutional neural networks, combinations of recurrent neural networks and convolutional neural networks, or others herein or herein. may include other types of neural networks or combinations or hybrids of types of neural networks described in the documents incorporated by reference. In embodiments, smart container designs may be optimized for cost, carbon emissions, speed, efficiency, performance, performance in specific environments (e.g., optimized for operation in arctic conditions), carrying capacity, safety, etc.

In another example, a set of machine learning models can be used to optimize the smart container path. For example, intelligence service 13004 may receive data from various sources described throughout this document and documents incorporated herein by reference, and generate a set of feature vectors based on the data received. there is. Intelligence service 13004 features trained machine learning models (e.g., using a combination of simulated and real-world data) to optimize a smart container route, e.g., based on a training dataset of results. Vectors can be input. In embodiments, intelligence service 13004 may include an input set of training data representative of market profitability optimization by a fleet of human experts and/or other systems or models. Data sources and feature vectors used for optimization of smart container routes include historical route data, order data, weather data, maritime data, traffic data, truck fleet data, rail data, news data, etc., to optimize smart container routes. It may include many types of shipping data described herein that may assist.

The foregoing examples are non-limiting examples and intelligence service 13004 may be used for any other suitable AI/machine learning related tasks performed for smart containers and shipping environments.

In embodiments, security system 13046 includes a framework that can be implemented at various levels of the disclosed systems. In these embodiments, instances of security system 13046 may be implemented at the system level, fleet or team level, or individual level. For example, at the system level, security system 13046 may provide security-related functionality for any communications and/or other interactions on behalf of system 13000 and/or with a smart container operating unit. In embodiments, the security system 13046 is implemented at the fleet level or team-level, where the security system is responsible for communicating and/or other interactions on behalf of a smart container team or fleet and/or with smart containers of a team or fleet. It can be configured to provide security-related functions. In embodiments, the security system 13046 implemented at the smart container level secures communications and/or other interactions on behalf of the smart container and/or with other smart containers, smart container teams, and/or systems 13000. It may be configured to provide related functionality.

In embodiments, security system 13046 may include an autonomous adaptive security module, an autonomous non-adaptive security module, and/or a passive security module. The autonomous adaptive security module may be configured to request intelligent tasks from the intelligence service 13004, whereby the adaptive security module evaluates security risks and determines actions based on the output of the intelligence service 13004. Utilizes artificial intelligence modules of intelligence service (13004). For example, the adaptive security module of a smart container fleet may be based on a set of data sources (e.g., weather data, video feeds, sensor data from the smart container and/or the environment, input from individual smart containers, etc.). One or more conditions associated with a smart container fleet can be monitored by receiving data from a set of data sources, such as monitoring routes for potentially hazardous conditions. In response to receiving the data, the adaptive security module may request an assessment (e.g., classification) of the environment from intelligence service 13004 regarding the security of the environment. In response, intelligence service 13004 may provide one or more classifications representing an assessment of the environment. The adaptive security module can then determine whether the evaluation requires action to be taken, and if so, what specific action to take. In some of these embodiments, the adaptive security module may use a rules-based approach to determine whether an assessment requires action and, if so, what action to take. Additionally or alternatively, the adaptive security module may be configured to provide a set of features (e.g., classifications, sensor readings from one or more smart containers, locations of the smart containers, objects detected in the environment and their locations). , and/or any other relevant features) may utilize a neural network that is trained on the action. In these embodiments, neural network module 13414 may receive features from an adaptive security module and/or set of intelligent service inputs 13470 and output a suggested action given the set of features. In some of these embodiments, intelligent service controller 13402 of intelligent service 13004 may accept or ignore decisions made by artificial intelligence module 13404. For example, analysis module 13408 may perform dynamic risk analysis and/or static risk analysis. Examples of dynamic risk analysis include real-time data-based analyzes (e.g. current weather patterns, current regulatory environment, current container port events, etc.) and/or risk analyzes specific to a particular cargo storage and/or transportation service order (e.g. (e.g., contractual risks, environmental risks, safety responsibilities, financial liabilities, etc.). Examples of static risk analysis may include, but are not limited to, operational risks (e.g., product design risks, manufacturing risks, quality control risks, etc.) and/or regulatory/compliance risks.

In embodiments, an autonomous adaptive security module may operate in an isolated manner (e.g., without communication with an external device or system) or in a connected manner (e.g., with communication with an external device or system). .

In embodiments, security system 13046 may include an autonomous, non-adaptive security module. In embodiments, the autonomous non-adaptive security module is configured to make security-related decisions autonomously (e.g., without human intervention) on behalf of the client. In embodiments, a non-adaptive security module performs logic-based security-related actions (e.g., risk mitigation actions) in response to detecting one or more specific sets of conditions. For example, the non-adaptive security module may, in response to detecting a particular set of conditions, lock the smart container, lock the digital twin of the smart container, power off the smart container, stop movement of the smart container, It may be configured to trigger actions such as starting a charge, sounding an alarm or siren, activating a strobe or light, sending a notification to another device or system, self-destructing, etc. In embodiments, non-adaptive security module 12284 responds to more easily diagnosable hazards, such as overheating conditions, movement or being pulled out of a geofenced area, detected internal leaks, low power conditions, low fluid levels, etc.

In embodiments, security system 13046 may include a passive security module. In embodiments, the passive security module is configured to enable a user to make decisions regarding security-related actions. In some of these embodiments, the passive security module is configured to receive notification of the assessed risk (e.g., from an adaptive security module, a non-adaptive security module, an intelligent client, etc.). In such embodiments, a human user may interface with the passive security module through a human interface, which may be provided through a user device (e.g., a mobile device, tablet, computing device, etc.).

In embodiments, fleet management system 13002 utilizes the features and capabilities of smart container system 13000 to achieve substantially optimized use of fleet resources by anticipating fleet resource demands and preparing those resources in advance of anticipated use. It can be done easily. In embodiments, resource demand anticipation may include coordinating delivery scheduling and maintenance activities to ensure that preventable outages due to lack of maintenance are avoided. Additionally or alternatively, resource demand projections may be based on alignment of detected fleet resource usage with information that supports prediction of cargo storage and/or transportation service orders, among other things. In embodiments, factors such as forecasted weather patterns, time of year, location, etc. may affect the likelihood of ordering specific cargo storage and/or transportation services. For example, high freight volumes are likely to occur during peak shipping periods from August to October, while freight volumes are likely to be low during the start of the year from January to March. Exemplary implementations for creating a fleet require predictions and handling these predictions follows a discussion of the components of the fleet management system 13002 and the associated smart container system 13000. As previously discussed, example components of smart container system 13000 may include communication management system 13010, remote control system 13012, and human interface system 13038.

In an embodiment, communications management system 13010 may include fleet management system 13002, as described herein, and elements thereof, intelligence services 13004, as described herein, and elements thereof, external data sources 13036, (e.g. Communication between system elements such as third party systems (e.g. via the Internet, etc.), smart container operating units, support systems and equipment, transportation resources (e.g. container ships, trucks, rail, etc.), human fleet resources, etc. (e.g. For example, it is configured to enable efficient and/or high-speed communication). Communications management system 13010 may include or provide access to one or more types of communications networks, such as wired, wireless, etc., which may support various data protocols such as Internet Protocol (IP). The communications management system may include, for example, intelligent services (e.g., through the fleet intelligence system resources described herein) that manage and control portions of the smart container fleet management system infrastructure associated with communications to ensure: It may include or have access to: Timely delivery of data collected by deployed smart container operating units to critical computational, analytical and/or data storage resources; Prioritized delivery of smart container configuration and operation commands; etc. Fleet Resource Management and Control In an embodiment, communications management system 13010 prioritizes the use of fleet security system communications of fleet communications resources over communications between fleet intelligence system components to support a high degree of security and integrity of fleet resources. You can get angry. Communications management system 13010 includes at least a network system 13202 that connects smart container system 13000 with external systems, deployed smart container operating units, and other network-connectable elements (e.g., fleet edge devices, etc.). Can provide and manage access to networking, including:

In embodiments, the capabilities of the communication management system 13010 may include, among other things, smart container system communication resources (e.g., network, radio) based on order execution plans, plan definitions, task definitions, smart container operating unit configuration, real-time task status, etc. may include context specifications, and/or adaptations of systems, data communication devices such as routers, etc.). Communications management system 13010 adaptation of fleet communications resources can be achieved by a variety of real-world conditions (e.g., weather, atmospheric conditions, container port traffic, container port and other facility structures, environments (e.g., land-to-submersible, underground). ), etc.). In embodiments, communications management system 13010 may obtain context from freight storage and/or transportation service orders that may facilitate anticipating the need for and type of adaptation during order execution. As an example of cargo storage and transportation service ordering context-based communication adaptation, an operation may begin at sea level and then involve action by underground teams. Communication resources suitable for use in these different task environments configured by the fleet composition system during work composition activities are adapted by communication management system 13010 for individual teams of smart containers as work progresses through the example environment. It can be controlled hostilely.

Freight storage and/or transportation service ordering criteria may directly invoke isolated operations. Alternatively, the circumstances of ordering cargo storage and/or transportation services may favor isolated operation (eg, operation within a foreign jurisdiction, etc.). Communication resources for the requested service can be adapted accordingly. By way of example, communication between teams of fleet resources assigned to juxtapose (e.g., co-locate at a point of origin) when performing a task may require a communications management system (e.g., as mentioned above when a team enters an external jurisdiction). (as indicated) may be configured by the fleet configuration system with additional encryption or with radio frequencies that defy conventional detection, which may facilitate activation when required by cargo storage and/or transportation service orders. In this additional embodiment of a fleet resource configuration, the communication management system 13010 may use communication resources (e.g., smart container operating unit wireless interfaces, communications proximate to isolated smart container operating units) to enforce such fleet configuration. infrastructure, etc.) can be detected and controlled. Another consideration for isolated operation is the use of low-energy short-range communications conditionally based on the deployment context, such as the expected location of team smart containers, such as when multiple smart container operating units are expected to be nearby. May include adaptive isolation communication protocols, such as allowing only Communication management system 13010 is responsible for configuring smart container operating units, selecting smart container units that meet cargo storage and/or transportation service ordering communication requirements, and deploying fleet communication resources (e.g., between teams and smart container operating units). It can assist the fleet configuration system with fleet configuration, such as configuration and assignment of relay devices (co-location), and other fleet and smart container configuration considerations. In this example of fleet configuration support, cargo storage and/or transportation service orders may indicate a preference for using specific smart container operating units. The fleet configuration system may query the communications control system regarding adaptive capabilities (e.g., the communications management system and/or specific fleet communications resources) to support the desired smart container operating units.

In an example of communication management adaptability capabilities to support various smart container operating unit communication configurations, communication management system 13010 may be configured to enable a first team of smart container operating units to perform operations in the same radio signal range as a first team of smart container operating units. By supporting the first team of smart container operating units to perform operations using different radio frequencies for wireless communication with the second team, the potential for cross-radio interference can be mitigated. Additionally, the communications management system 13010 may provide reliable network management through the use of redundancy, e.g., dual radio systems, automatic channel selection (e.g., local networking, cellular networking, mesh networking, long-range satellite networking, etc.). Communication can be provided. Fleet communication resources may include smart container operating units that act as network elements, such as when the smart container operating units are comprised of one or more mesh networks. Smart container operating units can visually utilize e.g. light sources (e.g. Morse code or binary transmissions), physical gestures, infrared signals, ant-based trails, etc. Communication may be facilitated in other ways, including through. Acoustic communications between smart containers (eg, non-human language encoded audio signaling), ultrasound, and other auditory-based technologies may be rendered as a form of communication between smart containers. Much like how co-located smart containers on different teams can use different radio frequency signals, co-located smart containers can use different auditory signaling to aid communication clarity between team members. .

In embodiments, communication management system 13010 may be configured as a plurality of independent communication systems configured to meet at least a corresponding portion of the fleet communication needs. In one example, communications management system 13010 includes a first communications system for communicating between elements within fleet management system 13002 (or any other fleet system, system, module, team, fleet segment, etc.), and intelligence. Service 13004 may be comprised of a second communication system for communicating between elements (or any other portion of a fleet system that may be separate from the first communication system), such that interference with any individual communication system interferes with other systems. It can be isolated from communication systems, thereby reducing the impact of communication problems throughout the system. Additionally, in this example, fleet management system 13002 and its components (e.g., work organization system 13018, etc.) may continue to communicate via the first communication system, even if an intelligent service such as a machine learning system (13004) Virtually all relevant fleet operational functions (remotely deployed fleet smart container operating units) may be compromised due to issues with the second communication system serving the intelligence service (13004). (including communication with others) can be performed. Additionally, communications management system 13010 may provide security features that affect the isolation and shunning of systems, sub-systems, system elements, communications systems, and other system resources that appear to be compromised by malware, etc. may include. Other independent communication systems include smart container-smart container communication systems, robot-smart container communication systems, human-smart container communication systems, emergency response communication systems, etc. Still other independent communication systems may be based on aspects such as confidentiality of information (eg, negotiations between a fleet management provider and shipper), fleet operation oversight, etc. In embodiments, communication management system 13010 may be configured to provide role-based (or other) access to different communication systems. For example, a fleet operations executive may be granted concurrent access to smart container operating units assigned to different tasks to perform fleet oversight functions.

In addition to and/or instead of separate communication systems, communication management system 13010 may provide redundancy (multiple frequency radios, etc.) to address exceptional conditions that may cause network damage, and emergency use. For this purpose, overriding operational communication channels may be required.

In an embodiment, communication management system 13010 may configure a fleet resource-specific (e.g., For example, individual smart containers can provide secure communications (operating units). Communications management system 13010 may additionally provide broadcast capabilities to support notifications, updates, alerts, and other services. Broadcast capabilities can be fleet-wide (e.g., notification to all fleet resources to observe daylight saving time), team-specific (e.g., notification of all team members' role changes), and updates on team members), task-specific (e.g., alerts on fleet resources assigned to a task that may include multiple smart container teams where work is pending), specific types of fleet resources (e.g. For example, it may be fleet resource type-specific, fleet support units, location-specific units (e.g., all units within a foreign jurisdiction), etc. to solve problems regarding smart container operating units).

In embodiments, communications management system 13010 may include, without limitation, a security system 13046, a network system 13202, and a variety of devices including artificial intelligence (AI) chipsets, data encoders, communications spectrum frequencies, etc. Task-specific communication elements may be used or managed in conjunction with other fleet management system features or services, including resources. Communications management system 13010 may operate in conjunction with security system 13046, such as by providing secure high-up-time access to fleets and associated communications resources. As an example, security system 13046 may utilize some of the configured communication channels (e.g., wired inter-computer links, wireless networks, etc.) that may be reserved by the communication management system for secure use. You can. That portion may include physically dedicated elements (eg, wired connections, wireless access points operating over a dedicated set of frequencies, etc.). In embodiments, providing dedicated wireless access includes prioritizing secure system access to existing wireless networks, such as by routing secure system data packets, streams, etc. ahead of non-secure system packets. can do. As another example, a communications management system may allocate communications devices with greater battery energy (higher charge) and/or a fixed power supply for secure system use, while lower power, lower power supplies, and/or power supplies are used for non-secure system use. Energy, and/or rechargeable devices may be allocated. Security system communications resource management and control is fleet-wide, job-specific, team-specific, deployment locale-based, and geolocation-based. It may be, etc.

Additional cooperative operations of communications management system 13010 and security system 13046 may be used to manage access by fleet resources to external resources (e.g., websites, etc.) as well as to manage access by fleet resources to external resources (e.g., websites, etc.). May include managing access by resources. Security system 13046 may deploy security agents, etc., to fleet resources based on the allocation and/or configuration of those resources. As an example, the firewall-type security functionality of security system 13046 may be deployed, among other things, at access points managed by a communications management system to connect separate task-specific communications systems.

In embodiments, communications management system 13010 may be configured to enrich fleet security capabilities and establish dynamic communications management functions that work in conjunction with fleet security capabilities to further reduce the likelihood of successful intrusion into a fleet communications system. For this purpose, it is possible to utilize the intelligence capabilities of fleet resources, such as resources with artificial intelligence capabilities (optionally provided by AI-specific chips and chip sets, etc.). By way of example, relying on AI-based functionality deployed across at least some of the fleet's resources (e.g., individual smart container operating units, etc.) (e.g., based on context and historical information representative of these environments, etc. ) can detect local environments with increased intrusion risk or other threats, so that a communications management system, optionally in conjunction with the security system 13046, can adapt fleet communications resources to reduce this risk.

Communications management system 13010 may facilitate and/or utilize control of use by others of network system 13202. As an example of management of network system 13202, communication management system 13010 may, for example, determine which resources utilize the network, how resources utilizing the network simultaneously may be coordinated, and limit network loading on those resources. By determining and/or controlling these, etc., the network system 13202 can be treated as a resource to be managed for use by fleet resources for communication.

In an embodiment, the smart container system includes a remote control system 13012 configured to provide a framework for remotely controlling smart container operating units and other external resources to complete cargo storage and/or transportation service orders. In embodiments, remote control system 13012 may manage the definition and use of control signals for remote operation of smart container operating units, fleet support units, external resources, etc. Smart container remote control enabled by remote control system 13012 may occur, for example, when a team supervising smart container is instructing one or more smart container team members to self-load a container vessel and Likewise, it may include the definition and management of a local smart container operating unit for smart container operating unit control signaling. Other examples of remote control signal management may include smart container-to-smart container fleet support signaling, intra-team smart container operating unit signaling, etc.

In embodiments, remote control system 13012 is configured to assist order execution system 13022 and provide a framework for remotely controlling smart container operating units and other external resources to complete tasks and/or operations. The remote control system can manage the definition and use of control signals for remote operation of smart container operating units, fleet support units, external resources, etc. Smart container remote control enabled by remote control system 13012 controls the smart container operating unit, for example, when a team supervising smart container directs one or more smart container team members to load and/or unload cargo. May include defining and managing local smart container operating units for signaling. Other examples of remote control signal management may include smart container-to-smart container fleet support signaling, smart container operating unit signaling within a team, etc. In embodiments, the remote control system may use resources of smart container system 13000, including, for example, communication management system 13010, security system 13046, and/or network system 13202, in some cases. access information, make decisions, and execute commands. A framework for remotely controlling a smart container operating unit may include a set of action-based standard rules, adaptive rules modified by situational awareness, emergency rules, exceptions, human decisions, ethical rules, fleet intelligence systems, etc. However, specialized, fall-over, or other communications required to handle a range of remote control requirements may facilitate delivery of remote control communications and/or communications from the use of remote control system 13012. This may be part of the communication management system 13010 that signals what to expect.

Remote control system 13012 may include local supervisor remote control initiators, human (local or remote) remote control initiators, automated fleet-based remote control initiators (e.g., fleet artificial intelligence systems, etc.), and ( Multiple initiators of remote control signals may be recognized, including third party remote control initiators (e.g., for law enforcement, etc.). Remote control signaling may include managing remote control signals to resources external to the fleet, such as fire and emergency response resources, infrastructure resources, third party smart container service providers, etc.

Fleet resources that can participate in remote control operations can vary in both implementation and protocols, such as older generation smart container operating units, human fleet resources, quantum computing elements, etc. Accordingly, the remote control system 13012 (in cooperation with the communications management system 13010) has multiple remote operation protocol (multi-protocol) capabilities to ensure that any two devices exchanging control signals can reliably do so. It can be composed of knowledge. In embodiments, multi-protocol capabilities may include handling and/or providing services as protocol-to-protocol conversion, remote control signal integration and interpretation, protocol normalization, etc. In embodiments, communications management system 13010 processes these protocols, either directly, as described above, via APIs, etc., or by being configured with such protocol handling capabilities (e.g., deployed with the protocol handling capabilities of remote control system 13012). You can use your abilities. In embodiments, remote control system 13012 (or equivalent functionality thereof integrated with communications management system 13010) may utilize digital twins and/or artificial intelligence, for example, to facilitate protocol conversion and/or adaptation. May rely on parts of intelligent services 13004, such as services. Accordingly, remote control system 13012 may provide real-time, on-demand protocol conversion optionally assisted by a fleet intelligence system. Remote control system 13012 may support fleet external remote control through ports configured for integration with external and/or third party remote control architectures. Remote control may be communicated via dedicated infrastructure and/or communication features (eg, short-range broadcast capabilities).

Remote control, such as control of a smart container operating unit, may be initiated, at least in part, by a human operator. In embodiments, smart container 13026 may encounter unexpected and/or unknown conditions during order execution (e.g., as may be illustratively reported by order execution system 13022), and the smart container The operating unit(s) may be entrusted to a human operator to remotely control them. Optionally, one or more intelligence service 13004 components, such as an artificial intelligence system, may be referenced for at least candidate remote control signals. In embodiments, an order execution plan may indicate that smart container operations, in predetermined operational tasks, should be guided by a human operator. When such a task is expected to occur in the work workflow (e.g., by a transport execution monitoring instance, such as a supervising smart container, etc.), the remote control system 13012 can activate the smart container with the appropriate human operator, smart container operator, etc. Can be called upon to supervise remote control connections between units, teams, team supervisors, etc., executing workflows that invoke human operator control.

In embodiments, remote control system 13012 may access a set of remote control signal sequences to remotely perform specific tasks. Remote control system 13012 may suggest one or more remote control signal sequences to a human operator and/or autonomous control system based on the context of the workflow being performed. In embodiments, the remote control system may, optionally with the assistance of other fleet resources (e.g., artificial intelligence systems, etc.), receive input from a human operator (e.g., "stop", "unload", etc.). commands) and generate a set of remote control signals to remotely control fleet resources, such as smart container operating units. Remote control signal sequences can be pre-configured to handle a range of real-time situations such as security breaches, equipment failures, etc. In addition to facilitating and/or managing remote control of a smart container operating unit, remote control signal sequences may be used for reconfiguration of deployed and/or allocated fleet resources for jobs, tasks, workflows, etc. A human operator (or automated system monitor-type application) can perform task roles, such as communicating remote control signals to one or more surviving members to communicate with the smart container operating unit configuration server to receive reconfiguration commands and reconfiguration data. and remote control signals communicated to surviving members of the team to coordinate actions.

Although generally described herein as a remote control signal, the remote control system 13012 transmits remote control signals at the fleet level, team level, smart container level, etc. to remote control commands (e.g., combinations of remote control signals, abstraction thereof). etc.), remote control can be facilitated. As an example of a remote control command function, remote control system 13012 may receive input from a human operator who wishes to command a smart container to raise a ramp onto a container ship, for example. In this example, the remote control system receives human operator remote control commands, adapts the commands to one or more different remote control signals for smart container 13026, generates corresponding remote control signals, and generates corresponding remote control signals by the human operator. It may ensure communication of these signals (e.g., via communication management system 13010 resources) to a smart container 13026 to be remotely controlled.

Smart container operating unit responsiveness to aggregated remote control signals (e.g., commands or set of commands) may be based on broad fleet intelligence capabilities, knowledge, priorities, goals, etc. In general, the use of system-based and/or smart container operating unit-based artificial intelligence capabilities supports broader independent decision-making capabilities for individual smart container operating units with greater context gravity.

In embodiments, remote control system 13012 may incorporate security features to prevent takeover, compromise, misuse, or interference with control of a remotely controlled smart container operating unit. Resources used by remote control system 13012 (e.g., data storage resources, computing resources, remote control system state data, etc.) may be configured using security features such as encoding, decoding, packetization, etc. there is. Additionally, the remote control system 13012 may allow a human operator to remotely control a smart container that is not directly involved in remote control signaling or operates independently of the remote control signal, e.g., autonomously, in cooperation with other smart container operating units, etc. May include and/or support control override capabilities to enable safe acquisition.

In embodiments, the smart container system 13000 allows a user to access the smart container system 13000 and/or an individual smart container operating unit (e.g., a user device, VR device, AR device, etc.) from a remote device (e.g., a user device, a VR device, an AR device, etc.). and a human interface system 13038 that provides a human interface that allows access (for remote control). In embodiments, human interface system 13038 may be used to enter cargo storage and/or transportation service orders (including any task-related parameters), manage fleet operations, manage fleet resources, manage fleet computing systems, software, and data structures (e.g., (e.g., system upgrades, etc.), human access to the smart container operating unit (e.g., for remote control of the smart container operating unit), augmented and/or virtual reality visualization of fleet operations, and (e.g., one Facilitates data extraction (for the creation and/or verification of smart contracts related to cargo storage and/or transportation service orders, etc.). As an example of use of human interface system 13038, a user may access status updates of a requested task via human interface system 13038. Users can use remote devices to observe smart container operating units performing tasks for requested tasks. In this example, human interface system 13038 interacts with other fleet components, such as order execution system 13022, to capture image capture resources (e.g., For example, you can direct a camera-based overhead drone).

In an embodiment, fleet management system 13002 includes work organization system 13018, fleet organization system 13020, resource provisioning system 13014, logistics system 13016, and order execution, monitoring, and reporting system 13022. (also referred to as “order execution system” 13022).

In embodiments, fleet management system 13002 may be configured to manage smart container teams, smart container fleets, smart containers, and/or support resources (e.g., edge devices, communication devices, container ships, cranes, additive manufacturing systems (e.g., It includes a resource provisioning system 13014 that manages provisioning resources for smart container operating units within the fleet, such as provisioning resources for smart container operating units (e.g., 3D printers, etc.). In embodiments, resources may include physical resources, digital resources, and/or consumable resources. Examples of physical resources include end effectors/manipulators, environmental shielding components, sensors and/or sensor systems, companion resources (e.g., drones, robots, container ships, trucks, rail systems). , cranes, lifts, etc.), hardware resources (e.g., special processing modules, data storage, networking modules, tethering modules, etc.), spare parts, human resources (e.g., technicians) , operators, etc.), power sources (e.g., generators, portable batteries), etc. Non-limiting examples of digital resources may include software, operating parameters, task-specific data sets, etc. Non-limiting examples of consumable resources may include fuel, packaging supplies, welding supplies, cleaning/cleaning supplies, etc.

In embodiments, resource provisioning system 13014 may support fleet-specific inventory, regional public inventory, rental/per-use fee-based resource inventories, and on-demand resource production systems ( Physical resources may be provided from an inventory of physical resources such as (for example, 3D printing of end effectors, etc.), third party inventory, etc.

In embodiments, resource provisioning system 13014 may operate cooperatively with other systems of the fleet operating system, such as fleet configuration systems, fleet resource scheduling and utilization systems, etc., to ensure that fleet resource provisioning rules are followed. there is. Physical resources to be provisioned may also include computing resources, such as on-smart container computing resources, smart container operating unit-local fleet-control computing resources, cloud/third-party based computing resources, computing and other modules and chips ( For example, for deployment with/within a smart container operating unit), etc. In some embodiments, fleet resource provisioning rules may be defined in governance standard libraries such that the resource provisioning system 13014 interfaces with an intelligent service to ensure that provisioned resources comply with the provisioning rules.

In an embodiment, digital resources to be provisioned by resource provisioning system 13014 may include pushing software/firmware updates (e.g., to update the onboard software of a smart container), resource access credentials, etc. ) (e.g., to access network resources such as task-specific smart container configuration data, etc.), on-smart container data may be provisioned through fleet configuration capabilities such as storage configuration/allocation/utilization data, etc. Use of provisioning system 13014 may include provisioning equipment, materials, software, data structures, etc., made and/or sourced specifically for a given cargo storage and/or transportation service order.

In embodiments, provisioning system 13014 may also operate in conjunction with contract systems, such as third party smart contract systems. In some embodiments, ordering freight storage and/or transportation services may reference or include a smart contract that may include and/or result in the configuration of an instance of provisioning system 13014 to comply with the request. As an example, provisioning system 13014 may receive smart contract terms that invoke provisioning constraints and/or guidance, such as from task configuration system 13018. Provisioning system 13014 may interpret these contract terms and thereby create a set of fleet and consumable resource provisioning constraints.

While the examples described above for resource provisioning system 13014 generally focus on provisioning related to order execution, resource provisioning system 13014 may also be used to provision fleet resources such as computing resources, fleet configuration systems, intelligent services, etc. Access to and/or execution of fleet elements may be further handled. In embodiments, provisioning of specific resources may be performed as part of a negotiation workflow for acceptance of cargo storage and/or transportation service orders. As an example, provisioning certain Intelligence Services (e.g., a fleet-level Intelligence Service) may result in higher rates for the shipper than other Intelligence Services (e.g., a Smart Container-level Intelligence Service may require a Smart Container Operating Unit (Only when placed in the field). As mentioned above and elsewhere in this specification, intelligent services can bring value to the system's fleet and work organization capabilities; Accordingly, provisioning such systems as part of negotiating cargo storage and/or transportation service orders may justify additional costs to the shipper. In some scenarios, prioritization of system resources, such as a fleet configuration system, may affect provisioning system 13014 functions.

In embodiments, fleet management system 13002 may, among other things, meet delivery requirements, maintain smart containers, maintain availability of fleet resources (smart container operating units, physical resources, etc.), and and a logistics system 13016 that handles logistics planning and execution for the pickup and delivery of items (e.g., replacement parts, end effectors, supplies, etc.). In embodiments, logistics system 13016 may utilize intelligence services, such as machine learning systems and/or artificial intelligence systems, to recommend logistics plans.

Logistics plans may refer to the workflow created to result in the delivery of a set of items to a specific location. In embodiments, logistics system 13016 may generate a logistics plan that utilizes fleet resources (e.g., smart containers, container ships, robots, trucks, cranes, railroads, etc.) to execute the logistics plan. In embodiments, fleet operating system 13002 may utilize (system-level) intelligence services 13004 to assist with logistics planning and decision making.

In an embodiment, fleet management system 13002 includes a maintenance management system 13028 that can be configured to schedule and perform maintenance on fleet resources, such as smart container operating units. Maintenance management system 13028 is responsible for field maintenance needs, including scheduled maintenance of fleet resources in the field to mitigate the impact on smart container operating unit utilization due to movement from deployed work sites to repair garages, and Requests can be handled. Maintenance management system 13028 may also coordinate maintenance and repair operations at repair garages, etc. In embodiments, maintenance management system 13028 may include, provide access to, and/or integrate with mobile maintenance vehicles, spare parts garages, third party maintenance service providers, etc. You can. In embodiments, maintenance requirements for fleet resources housed in storage areas such as warehouses, remote inventory garages, etc. can be proactively addressed to smart containers so that the smart containers are less likely to require maintenance during deployment. May be evaluated by maintenance management system 13028 for pre-scheduled maintenance, such as when maintenance activities are approaching.

In embodiments, maintenance management system 13028 provides the status of fleet resources, such as smart container operating units, on a scheduled basis or in response to queries about smart container operating unit status by maintenance management system 13028, etc. This can be monitored through resource status reporting. In an embodiment, maintenance management system 13028 may be configured to include a smart container operating unit that signals that the smart container operating unit is experiencing reduced power output, and that reports exposure to certain ambient conditions (e.g., excessive heat). Smart container operating unit, smart container operating unit reporting a leak containing liquid cargo requiring cleaning, lack of heartbeat signal from smart container operating unit to smart container health monitor resource Smart container operating unit communications can be monitored for indications of potential service conditions, such as: Additionally, the maintenance management system 13028 may monitor information within a smart container data store that stores smart container operating unit status information, activates self-test operation modes, collects data that provides an indication of smart container maintenance needs, etc. Probes can be placed within smart container operation and/or supervision software that can perform maintenance management functions on the same smart container operating unit. Maintenance management system 13028 may also include maintenance robots that can be deployed with smart containers in teams of smart container operating units to perform requested tasks.

Maintenance management system 13028 utilizes human/operator input (e.g., a human observer may indicate that a smart container operating unit appears to be behaving erratically), smart container process of maintenance activities. Artificial intelligence to predict maintenance instances for automation, scheduling, and machine learning to help identify new opportunities to perform scheduling and maintenance (e.g., cooling systems in smart containers before operating in high-temperature environments). Configured to utilize various system services and capabilities to schedule and perform maintenance, including analyzing the performance of smart container operating units maintained for certain conditions before operating under those conditions, such as upgrading It can be. In embodiments, maintenance management system 13028 may receive maintenance-related input. In embodiments, candidate sources of maintenance-related input include human operators/observers, maintenance scheduling services, third-party service providers, smart container production vendors, and parts providers for scheduling maintenance. can do. Maintenance management system 13028 may also implement business rules (e.g., determined by regulatory agencies, by shipper, team, fleet, etc.) to determine appropriate maintenance workflows, service actions, required parts, etc. established rules, etc.), association tables, data sets, databases, and/or maintenance management libraries. In embodiments, maintenance activities may be assigned by a maintenance management system to fleet resources, such as maintenance smart containers, human technicians, third-party service providers, etc.

In embodiments, deployed smart container operating units may be configured with one or more maintenance protocols to perform self-maintenance, such as self-cleaning, correcting end-effector operations, etc., among other things. Self-maintenance reduces the ability to respond to the detection of damaged smart container operating unit features, such as defective 3D printing systems or defective systems for securing cargo (e.g. steel strapping, polyester strapping, dunnage bags, etc.). It can be included without limitation. A deployed smart container operating unit may determine that a capability has been compromised and, optionally with the assistance of maintenance management system 13028, other smart containers such that the compromised capability can be resolved when time allows rather than causing a delay in completion of delivery. Allocation with containers can be switched. Additionally, smart container operating unit intelligence (e.g., on-smart container AI, etc.) may predict damage to a smart container capability, for example, based on time-to-failure data for the smart container capability.

In an embodiment, maintenance management system 13028 may use intelligence services 13004 (e.g., system level intelligence services (e.g., system level intelligence services) to predict when maintenance may be performed on a smart container operating unit and/or its components. 13004)) can be used. In some of these embodiments, maintenance management system 13028 may request a digital twin of a smart container operating unit from intelligence service 13004. At this time, the digital twin can reflect the current state of the smart container operating unit and analyze the digital twin of the smart container operating unit to determine whether maintenance is required for the smart container operating unit. Additionally or alternatively, the digital twin service of intelligence service 13004 may run one or more simulations involving smart container operating units to predict when maintenance may be required. In some of these embodiments, the output of a smart container operating unit's digital twin may be analyzed (e.g., using machine learning predictive models or neural networks) to predict if/when maintenance may be needed.

In an embodiment, fleet management system 13002 includes work organization system 13018. In embodiments, a delivery configuration system receives an order for cargo storage and/or transportation services, such as from a customer booking a smart container delivery service. In embodiments, a cargo storage and/or transportation service order may represent a set of cargo storage and/or transportation service order parameters. Non-limiting examples of cargo storage and/or transportation service ordering parameters include: timing requirements, origin and destination of shipment (e.g., region, address, coordinates, etc.), number of smart containers required, type of container required (e.g. For example, tank container, bulk container, 20-ft standard container, 40-ft high-cube container, etc.), container utilization requirements (e.g., full container (FCL) vs. shared container (LCL)), cargo description (e.g. (e.g., number of packages, total volume, or total weight), whether the shipment contains personal effects (person effects), and other necessary tasks (e.g., inspection tasks, packaging tasks, unpackaging tasks, unloading tasks, loading tasks). , 3D printing tasks, growth tasks, assembly tasks, monitoring tasks, etc.), pricing information, and any other suitable parameters. In embodiments, cargo storage and/or transportation service ordering parameters are required that may indicate what type of smart container operating unit is needed and/or its capabilities. These and other freight storage and/or transportation service ordering details are described elsewhere herein.

In embodiments, quantum optimization may be enabled by quantum computing service 13008, which may optimize allocations across fleet resources, such as smart container operating units. Quantum computing service 13008 can further optimize routing (logical, physical, and electronic) associated with smart container fleets, shipping, teams, communications, logistics, etc. Additionally or alternatively, in some embodiments, quantum computing service 13008 may perform various fleet functions including energy consumption, compute capacity and utilization, infrastructure resource planning, engagement and utilization, risk management, compute storage capacity, etc. It can be used to optimize combinations of smart container resources and other resources. Quantum computing service 13008 may also optimize smart container design, optimize smart container service pricing, and optimize smart container charging (e.g., to ensure that smart containers with solar panels receive sufficient levels of sunlight). can be used to optimize its route, etc.

In embodiments, the work configuration system 13018 and other fleet resources (e.g., fleet configuration, system intelligence, smart container behavior, etc.) can, among other things, learn from failures as well as execute tasks, workflows, and orders. Benefit from the use of deep learning techniques for plan optimization. In these embodiments, work organization system 13018 may request deep learning services from system-level intelligence service 13004, which may be a set of features, including features extracted from freight storage and/or transportation service orders. Utilize neural networks and/or other machine learning models to determine task configurations based on In these embodiments, artificial intelligence services may be configured to learn task workflows, job configurations, etc.

In embodiments, task configuration, fleet configuration (which may include smart container configuration), and/or order execution may be used to improve fleet functionality, performance, and results through the use of local context adaptive task assignment, execution, resource routing, etc. Additional improvements can be made. This adaptive capability can be further enabled through peer-to-peer based communication (e.g., smart container operating units within a team) that quickly and efficiently reveals the context of work activities.

In embodiments, artificial intelligence for automation of smart container allocation and execution (e.g., smart container process automation through learning) may be used in a fleet management system (13002), for example, to learn smart container allocation from human operator allocation activities. ) and intelligent services 13004 may function in cooperation with elements of the smart container system 13000. Other learnings that artificial intelligence systems can produce in the context of smart container fleet configuration and behavior include task completion, completion time, percentage of intact shipments delivered, cost of completion, quality of completion, ROI on resources, resource utilization, etc. can be based on outcome measures of success.

These and other task organization details, including the operational flows of task organization system 13018, are shown and described in the relevant figures herein.

In embodiments, fleet management system 13002 may operate fleet resources (e.g., smart container operations) to satisfy cargo storage and/or transportation service orders from multiple concurrent and/or overlapping cargo storage and/or transportation service orders. and a fleet and smart container configuration system 13020 (also referred to as fleet configuration system 13020) that may operate in cooperation with the work configuration system 13018 to determine the configuration of units, teams, etc. Fleet configuration system 13020 is responsible for ordering cargo storage and/or transportation services, required tasks, budgets, timelines, availability of smart containers or smart container types, availability of container ships or other transportation modes, traffic at container terminals or ports, and/or determine fleet and smart container configurations based on other appropriate considerations. In some embodiments, fleet configuration system 13020 may utilize system level intelligence service 13004 to determine fleet and/or smart container configuration. In some of these embodiments, the intelligence request may include a proposed fleet configuration and other related data (e.g., cost constraints, cargo type, origin location, destination location, route environment, etc.). In response, intelligence service 13004 may output a proposed fleet configuration. Additional details of the fleet configuration system 13020 are described and shown in figures elsewhere herein.

In embodiments, fleet management system 13002 may include an order execution, monitoring, and reporting system 13022 (also referred to as order execution system 13022). The order fulfillment system 13022 is responsible for system functions such as logistics for smart container and fleet resource delivery, data collection, cataloging, library management, and data processing system 13024 allocation to facilitate data processing activities for order execution. An order execution plan may be received from work organization system 13018, which processes by coordinating activities. In general, order execution system 13022 may be configured other than that configured by task configuration system 13018, such as compute, storage, bandwidth, etc., as may be determined and/or defined by order execution plan useful for executing an order execution plan. Work can begin by committing and managing resources, including:

In embodiments, order fulfillment system 13022 may facilitate compliance with reporting requirements associated with order fulfillment (e.g., shipment-specific, fleet-specific, compliance-related reporting, etc.). In embodiments, reporting may include collecting data (e.g., from smart container operating units, sensor systems, user devices, databases, etc.), processing data, and preparing feedback for use of order execution data by task and fleet configuration systems, etc. may include. In embodiments, the order execution system 13022 may be configured with a maintenance management system 13028, a resource provisioning system 13014, and a communication management system 13010 that facilitate communication between smart container operating units, teams, fleets, etc. It may be assisted by other system capabilities to transmit, process, store and manage data affecting order execution. These and other fleet and external resources determine the operational behavior of the requested task, e.g., which communication resources have communications management system 13010 in reserve and/or assigned to the requested task, smart container operating units, and tasks. Information may be provided to the order execution system 13022 to facilitate servicing and/or maintenance requirements for other resources used in execution, changes to resource provisioning that occur after operation of the job begins, etc.

In an embodiment, the order execution system 13022 may, for example, determine whether an order execution system 13022 is experiencing an on-the-job condition (e.g., excessive container port traffic, unexpected ground conditions due to excessive rain, etc.). Identifying bottlenecks can further facilitate the evaluation and modification of order execution plans during job execution.

In embodiments, order fulfillment system 13022 may perform various data pipeline functions during execution of a job. In embodiments, data pipeline functions may include optimizing the use of pre-configured sensor and detection packages combining sensor selection, sensing, information collection, preprocessing, routing, integration, processing, etc., among others. . In embodiments, sensor and detection packages may be activated by order execution system 13022 when their use is indicated to provide scope for monitoring/reporting activities. Examples of other data pipeline features include optimizing on-smart container storage, selective sensor data filtering for reduced impact on communications bandwidth (e.g., reducing the need for wireless network utilization), and exceptional condition detection. and pipeline adaptation/data filtering, etc.

In embodiments, order execution system 13022 may monitor smart container power demand during order execution and resolve if necessary. In these embodiments, order execution system 13022 may provide battery charging capacity across multiple smart container operating units to meet work task and workflow requirements, such as queues of tasks that must not be interrupted. (or other energy source levels, such as fuel levels). In embodiments, smart container power demand management may include routing fleets, teams, and individual smart container operating units to complete tasks with reduced delays in overall productivity with integrated smart container charging activities. Additional details of the functionality and operation of order execution system 13022 are described throughout this disclosure.

In embodiments, smart container functionality, including during order execution, can be implemented, for example, through the deployment and use of optionally automated smart container 3D printing and production capabilities for last-mile customization of products. For example, it can be combined with 3D printing services and systems to enable agile, remote, and flexible manufacturing on an on-demand basis.

In embodiments, order execution system 13022 may execute, deploy, and/or interface with a set of smart contracts that monitor and report to smart container operating unit 13040. In embodiments, a powerful distributed data system, such as a distributed ledger (e.g., a public or private blockchain), can track and enhance smart container fleets and/or smart container activity, as well as reduce the cost of smart container resource usage to the parties involved. Can be used for allocation. In some of these embodiments, distributed ledger nodes store and execute smart contracts. In embodiments, smart contracts may be configured to monitor cargo storage and/or transportation service orders, order execution, resource usage, etc. For example, in some embodiments, smart container operating units may be configured to provide a smart contract with evidence of completion of a task (e.g., delivery of a shipment), so that the smart contract can take action in response to completed tasks. (e.g. payments, records, etc.) can be triggered. In another example, a smart container operating unit may be configured to report location data, sensor data, status data (e.g., fill level, cargo status and/or status, etc.), and/or other appropriate data, thereby Contracts can be configured to trigger specific actions based on received data.

In embodiments, smart container system 13000 may include, among other things, scalable computational capabilities, data management capabilities (e.g., data caching, storage allocation and management) for any smart container freight service operations and/or intelligent resources. etc.), a data processing system 13024 that can provide access to and control of fleet and/or task-related data stores, such as libraries, fleet resource inventory control and management data structures, etc.

In embodiments, fleet management system 13002 may provide support for fulfilling cargo storage and/or transportation service orders. For example, components of the fleet management system 13002 may manage fleet resources (e.g., smart container operating units, physical modules, etc.) to fulfill cargo storage and/or transportation service orders, such as the timing of order execution, etc. and/or supporting devices) are provided in an efficient manner. For example, in some embodiments, fleet management system 13002 may provide delivery of fleet resources and/or maintenance tasks to ensure that fleet resources are allocated in an efficient manner without significantly impacting task completion times. “Just-in-time” strategies can be used to facilitate. In some of these embodiments, fleet management system 13002 may be configured to: predict fleet resource demands corresponding to various cargo storage and/or transportation service orders; Or, intelligent services can be leveraged to predict order execution plans to prepare for maintenance.

In embodiments, order fulfillment system 13022 may predict task-related resource demands in a task-specific manner to predict when certain resources will be needed for a particular task. For example, order execution system 13022 (working in combination with an intelligence service) may generate a schedule of ongoing and/or upcoming work for a particular freight storage and/or transportation service order, and in response thereto: You can determine when specific fleet resources are likely to be needed and/or available. Additionally or alternatively, order execution system 13022 may estimate task-related resources for a particular task in any other suitable manner. For example, forecasting resource demands may include patterns of fleet resource demands derived from a shipper's cargo storage and/or transportation service order history; the shipper's resource usage history from N previous operations performed on the shipper; The timing of cargo storage and/or transportation service orders may be determined (e.g., orders are typically received on a Thursday for jobs to begin the following Monday), etc. Business relationships between entities may form the basis for predicting the shipper/buyer's fleet resource demands and timing based on actions involving the supplier/seller/consumer's freight storage and/or transportation service orders.

In embodiments, many other factors may affect fleet resource demand forecasts, such as weather forecasts and seasonal effects. Fleet resource demand forecasts may also be activated by events outside of the core cargo storage and/or transportation service ordering process, such as natural disasters, accidents/emergencies, pandemics, etc. In other examples, other sources of information that can affect the projection of fleet resource demands include business objectives such as reducing or increasing spending near the end of a financial reporting period (e.g., fiscal quarter, year, etc.) and purposes. An indication that the target shipper intends to reduce costs in the final weeks or months of the fiscal reporting period may be an indication that fleet resources normally allocated by the target shipper to cargo storage and/or transportation service orders may be diverted from maintenance, upgrades, or other costs to other shippers. It may indicate that it will be available for other actions, such as allocation. In embodiments, fleet goals or objectives may also influence fleet resource expectations and corresponding preparation activities, etc. One such example is the required upgrade of a class of smart containers. In anticipation of the need to reserve smart containers in this class, fleet configuration functions can assign alternative smart container types that can be reconfigured to meet the requirements of the reserved smart container class for the duration of the upgrade activity. .

In embodiments, projections of fleet resource demands may be determined through the use of smart container systems 13000, such as intelligent services 13004 and fleet management systems 13002. For example, in some embodiments, intelligence service 13004 may provide weather forecasts, public activity calendars, freight storage and/or transportation service ordering data (e.g., timing, operating parameters, other freight storage and/or or relationships to transportation service orders, etc.), social media activity, government activity, and/or legislation, etc., may analyze sources of data that may affect fleet resource demands. In this example, intelligence service 13004, operating in conjunction with fleet management system 13002, can predict fleet resource demand based on analysis of different data sources (e.g., using a neural network, etc.). In these embodiments, intelligence service 13004 can process data from different data sources and determine the likelihood of fleet resource demands across a variety of factors.

In embodiments, smart container system 13000 may use external data sources 13036 to perform various system functions, including task configuration, fleet configuration, task negotiation (e.g., via a smart contract facility), order execution, etc. You can interface with Examples of external data sources for use by the system include value chain entities (e.g., third parties paying for delivery services, etc.), teaming and/or executing requested tasks, smart contracts, etc. Includes ERPs (enterprise resource planning systems) that can provide a working context for Other external data sources may include third-party sensor systems (e.g., GPS data, value chain logistics data, etc.) as well as third-party data streams (e.g., weather, traffic, electricity rates, etc.). there is.

In some embodiments, smart container system 13000 may support the use of smart contracts in connection with ordering cargo storage and/or transportation services, performing tasks, allocating resources, etc. In embodiments, cargo storage and/or transportation service orders may be routed through a smart contract handler that captures task requirements, requester goals and objectives, and fleet order execution constraints into a dynamic smart contract. In some embodiments, a smart contract may transfer a smart contract from a first location (e.g., an origin location, container port or terminal, temporary storage/service location) to a second location (e.g., a destination location, container port or terminal, etc.). It can be used across smart container fleet management systems to handle all manner of fleet operations, such as managing negotiated routing of containers. As a further example, a smart contract may be executed as a control over the bidding system for smart container time/task utilization. As another example, a smart contract may monitor certain activities (e.g., task-related activities, etc.) related to cargo storage and/or transportation service ordering. Smart contracts rely on and/or benefit from access to fleet system data (e.g. route progress, sensor data, etc.) to trigger actions defined by the smart contract, such as payment upon completion of delivery of smart container cargo. can be obtained. The smart container system 13000 is a fleet storage system that contains fleet data through application programming interfaces (APIs), infrastructure elements such as sensor networks, edge computing systems, etc., to update states related to smart contract conditions. Can provide access to resources.

In an embodiment, work organization system 13018 and fleet organization system 13020 collectively generate an order execution plan in accordance with some embodiments of the present disclosure. In embodiments, an order execution plan may define a set of smart container paths and/or tasks to be performed upon completion of a requested task, and may further define a configuration of a fleet of smart container operating units to complete the task. . In embodiments, an order execution plan may include task definition (which may include route definition), workflow definition, fleet configuration (which may include smart container configuration of individual smart containers), team assignment, and container port site details. It may include reference to (or integration of) context information such as the like. In an embodiment, job configuration system 13018 receives an order defining a task to be performed, and task configuration system 13018 defines a task definition, each defining a delivery service and/or task to be performed by the smart container upon completion of the task. A set of can be determined. In an embodiment, task organization system 13018 further defines a set of workflow definitions. Workflow definitions define at least one order in which tasks are performed in completion of a project and/or task, including any loops, iterations, triggering conditions, etc. In embodiments, task organization system 13018 may determine workflows based on task definitions that include tasks. Task composition system 13018 may utilize libraries of pre-configured workflows to complete specific tasks. Additionally or alternatively, work organization system 13018 may utilize intelligence services 13004 to obtain initial workflow definitions for tasks and/or projects that are part of a larger job. In some embodiments, a human may construct the initial workflow definition and/or provide input used to determine the initial workflow definition. In embodiments, work configuration system 13018 may be configured to exchange information to develop a smart container fleet order execution plan and/or utilize one or more services of a smart container fleet order execution plan with one of the smart container systems 13000. It can interface with the above components. For example, work configuration system 13018 interfaces with data processing system 13024, smart container, smart container configuration library of fleet, project and task related information, fleet level intelligence service 13004, fleet configuration system 13020, etc. can do.

In embodiments, work configuration system 13018 may include a plurality of systems that perform order execution plan preparation functions by processing information received on freight storage and/or transportation service orders. In embodiments, the systems of task organization system 13018 may include a task parsing system, a task definition system, a workflow definition system, and a workflow simulation system. In the illustrated example, the work configuration system 13018 systems work in combination to generate an order execution plan that is used to define a set of smart container operating unit assignments. In embodiments, smart container operating unit assignments may supplement or be integrated with an order execution plan and may identify specific smart container teams and/or smart containers assigned to each task. For example, a smart container operating unit assignment may define specific tasks, and for each task, a specific smart container team with a smart container unique identifier and/or a team identifier assigned to the task. Smart containers can be identified. In embodiments, smart container operating unit assignments may be created by job configuration system 13018 and/or fleet configuration system 13020.

In embodiments, a task parsing system receives and parses cargo storage and/or transportation service orders to ultimately determine task definitions, project definition(s), task definitions, workflow definitions, fleet configurations, and smart container configurations. Determine the set of cargo storage and/or transportation service ordering parameters to be used. In embodiments, the task parsing system may be configured to receive input from an operator to configure, adapt, or otherwise facilitate parsing of freight storage and/or transportation service orders from a user via a user interface, such as human interface system 13038. May receive orders for cargo storage and/or transportation services. Additionally or alternatively, the task parsing system may receive an order for cargo storage and/or transportation services from a client device associated with the requesting organization.

In embodiments, the task parsing system may consist of a collection facility for receiving electronic versions of task descriptions and related documents, such as GPS data, smart contract data and/or due dates, links to them, etc. The collection facility can parse documents for keywords, references to activities, etc., which may be useful in determining work requirements. Additionally, keywords in the collected work content, such as weight terms, volume terms, path environment terms, etc., can be useful by work configuration system 13018 elements by providing insight into the type(s) of smart containers needed and their configurations. It can be applied easily. As an example, a keyword suggesting content to be moved weighs 25 tons, suggesting a smart container transport device/team with at least that moving capacity.

In embodiments, a task parsing system may incorporate machine learning capabilities (e.g., which may be provided by intelligence services 13004) to improve techniques for parsing task content, which may include descriptive data, and/ Or you can use In addition to machine-based learning from human-generated feedback on work content parsing results, learning can be performed on other work content parsing actions (e.g., previous freight storage and/or transportation service orders), common and specialized knowledge bases ( It may be based on experience with, for example, technical dictionaries), expert humans, etc.

In embodiments, parsing work content may include automated parsing of structured and unstructured text. In some embodiments, a task parsing system may be configured to identify (and selectively disassemble) missing/unclear data and eligible task content data (collectively referred to as “insufficient information”). In response to identifying insufficient information, the task parsing system may generate and provide a request to a human operator via a user interface for clarification of the insufficient information. This request may identify specific inputs from the user to provide, so that the request identifies disambiguation content that was initially missing or unclear. Additionally or alternatively, parsing system 13214 may be configured to retrieve information from a previous freight storage and/or transportation service order such that disambiguation content may be obtained using previous freight storage and/or transportation service order information and context from the request. Disambiguation content may be determined (e.g., via a query) from a library 13044 that holds the data. If the parsing operation cannot determine disambiguation content, parsing system 13214 may generate a request to disambiguate content as discussed above.

In embodiments, various job description information may be provided to, determined, and/or extracted from job configuration system 13018. Examples of freight storage and/or transportation service ordering parameters include origin location and destination location information and/or other physical locations along the route; weight requirements; volume requirements; cargo description; smart container type(s); Number of smart containers required; May include, but are not limited to, transportation modes (container ships, trucks, rail, self-driving smart containers, hyperloop, etc.). 3D printing requirements (such as last mile customization); 3D CAD models or digital data of environmental layouts such as ships, shipping container ports, shipping container storage facilities, etc. may be available or completed as part of the initial task scoping, task prioritization and Can be used to automatically provide workflow routing, smart container selection, supervision requirements, etc.; operating environment (such as along a route) including temperature, hazard statement(s), terrain, weather, etc.; Deliverables such as data, reports, analytics, etc.; Customer interfaces for data exchange such as network interfaces, APIs, security; Communication network availability such as land lines, 4G, 5G, Wi-Fi, private networks, satellite, connectivity constraints, etc.; budget constraints; Timing requirements including scheduling against port availability, scheduling against vessel availability, earliest start time, final end time, activity rate such as number of smart containers active at any given time, etc. Other examples of task description information that can be handled by the task parsing system include smart contract terms, certification levels of smart container operating software for deployed smart containers, insurance regulations, regulatory requirements (e.g., customs requirements) sites), site access requirements (e.g., a particular container port can only be accessed when a human is present or through coordination with humans present on the site), tasks, activities, workflows, or overall operations. May contain contract-related information, such as conditions for assigning a proxy.

In an embodiment, the task organization system 13018 system (e.g., task parsing system, task definition system, and workflow definition system) may combine smart container automation task content with other task content (e.g., cost, payment, financial, etc.). See library 13044 to identify content and structural filters to distinguish from pre-configured candidate tasks, workflows, and/or completed work configurations that substantially meet the requirements of a freight storage and/or transportation service order. can do. In embodiments, library 13044 or another task configuration library may facilitate mapping representations of task content with target terms that represent smart container automation. As an example of the use of automated tasks from library 13044, a requested cargo transportation service may include a requirement to measure the temperature inside the smart container using a set of temperature sensors. The task parsing system may identify temperature measurement requirements and, in response, the task definition system may provide a library (i.e., a library) that meets the requirements of that portion of the freight storage and/or transportation service order, which may be used to define an order execution plan. 13044) can identify an automated measurement task for measuring the temperature of a container. If the job configuration system 13018 determines that an appropriate job configuration is available (e.g., from library 13044), such as if ordered freight storage and/or transportation services have previously been requested, the job configuration system (13018) may use previous task configurations corresponding to previously requested tasks as proposed task configurations for further verification with current fleet standards, etc. For example, intelligence service 13004 may support a proposed working configuration (e.g., one or more including, without limitation, machine learning services) against a set of governance standards to ensure that the proposed working configuration is consistent with the standards. intelligent services) can be analyzed. Intelligence service 13004 may perform other intelligence-based tasks for the proposed task configuration.

In some scenarios, task organization system 13018 may determine that one or more tasks, workflows, routines, etc. do not have an appropriate counterpart in library 13044. In this scenario, the task parsing system collects smart container fleet-focused requirements (e.g., task definition parameters, smart container configuration parameters, proposed You can create a data set containing task sequence, etc.). In embodiments, the task parsing system may rely on intelligence service 13004 for suggestions of task requirements, including combinations of tasks that, when selectively adapted, can satisfy such requirements.

In embodiments, the operational parsing system may include and/or interface with analysis modules/governance libraries of intelligence service 13004 of system 13000. A task parsing system can leverage governance-based analytics by providing candidate smart container automation portions of task content (e.g., terms, etc.) for processing. Intelligence service 13004, in response to the provided portion of work content, provides one or more of the safety standards and/or one or more of the operating standards to be applied during preparation of the order execution plan by the work configuration system 13018 and/ Or it can be displayed.

In an embodiment, the job parsing system allows the job composition system 13018 to define smart container tasks, configure fleet resources, define workflows, simulate workflows, generate order execution plans, etc. It may include a job requirements module that creates a set of cargo storage and/or transportation service order instance specific requirements for use. In embodiments, a set of cargo storage and/or transportation service order instance specific requirements may include a candidate portion of work content representing smart container automation (e.g., a term representing a smart container task), one or more inputs from a user interface ( at least one of the following: (e.g., disambiguation of terms), safety and behavioral standards (e.g., from a governance layer), and recommended smart container tasks and related context information (e.g., provided by a fleet intelligence service). It can be decided based on .

In embodiments, the task parsing system may use content filters and/or to identify structural elements within the task content that may represent one or more of tasks, sub-tasks, task sequencing, task dependencies, task requirements, etc. Structural filters can be applied. In embodiments, detected structural elements may facilitate selection and configuration of smart container operating units, for example, by fleet configuration system 13020. In one example, a structural element that distinguishes a set of tasks can be used by a fleet configuration system to avoid assigning the same smart container operating unit to tasks within and outside the set of tasks described by the structural element.

In embodiments, the job parsing system may integrate and/or utilize a freight storage and/or transportation service order configuration agent/expert system that may be configured to facilitate the development of job description parsing capabilities.

In embodiments, a task definition system may organize work data into task definitions (e.g., individual smart container tasks or tasks performed by smart container teams). The task definition system may further coordinate other systems in the task composition system 13018, such as the workflow simulation system, to optimize task definitions.

In embodiments, the task definition system may refine the task data compiled by the task parsing system to facilitate defining the individual actions of one or more smart container operating units within a fleet of smart containers in performing the requested tasks. . Defining tasks includes smart containers, shipping container ports, container ships, railways, trucks, hyperloops, smart container types, smart container features, and smart container configurations that can perform the defined task. It can be based on information about people. In embodiments, the task definition system may further provide information in task definitions to facilitate fleet configuration system 13020 in determining the use of smart containers for each defined task. In embodiments, the task definition system may reference library 13044, intelligence service 13004, or other system-specific or accessible resources when making task suggestions.

When the task definition system defines the tasks of a job, the task definitions may be cataloged and stored for future use, such as in library 13044. In some embodiments, the task definition system may adapt a task definition from a previously cataloged task definition (e.g., to adapt a task for a particular type of environment or its specific conditions). Adapting the definition). In these embodiments, the task definition system may catalog the derived task definition within library 13044 with adaptation instructions. In some embodiments, a task definition cataloged in library 13044 may be associated with an already cataloged task definition and/or replace an already cataloged task definition and cataloged as a subtask of an existing task. It can be done, etc. In general, a task definition may include associated tasks, serialized tasks, nested tasks, etc.

Information about the task may be stored in library 13044 for future use; Accordingly, the task definition system can access library 13044 to retrieve information about tasks, smart containers, fleets, etc. In an example embodiment of self-stacking smart containers on a container ship, information accessible through library 13044 may include, for example, methods to access information regarding the physical layout of the container ship. The task definition system may also use the library to update information, for example, by adding one or more tasks to the list of tasks for a self-stacked task, results from optimizations of the task definition performed by the order execution system, etc. (13044) can be accessed.

Optimization features of the task definition system are described below in relation to feedback from other elements of the task composition system 13018, such as the workflow simulation system.

Task definitions may be created and provided to other elements of the work composition system 13018, such as the workflow definition system and the fleet composition system proxy. In embodiments, a fleet may provide task definitions (which may include path definitions and other pertinent information) to the fleet configuration system 13020. In one example, a fleet configuration system proxy is configured to perform tasks (indicated in task descriptions) based on fleet configuration and fleet resource inventory and allocation data associated with the requested task (e.g., based on geography, timing, etc.). The set of candidate smart containers can be narrowed to a specific smart container type (and optionally a specific smart container within the fleet). The fleet configuration system proxy may process task definitions, which may include smart container identification information (e.g., smart container type, etc.) to align the fleet's resources with associated task information. In one example, a fleet configuration system proxy is configured to perform fleet resource allocation, scheduling, etc. in support of at least some of the objectives of cargo storage and/or transportation service orders being processed through task configuration system 13018. Data suitable for use by fleet operation elements such as the resource provisioning system 13014 may be generated. The fleet configuration system proxy may use fleet configuration modeling to determine candidate fleet configurations that meet job requirements. Modeling can be useful in determining impacts on fleet resources that can be considered during fleet configuration functions, resource allocation, etc. In embodiments, fleet configuration modeling may include the use of system intelligence service resources, such as machine learning, artificial intelligence, etc., in determining one or more preferred fleet configurations that also meet one or more job description requirements. Fleet configuration system 13020 is described in more detail elsewhere in this disclosure.

In an embodiment, work organization system 13018 receives task definitions from task definition system, receives fleet configuration information from fleet organization system 13020, and performs task sequencing (e.g., timing of deliverables and/or tasks). and a workflow definition system for receiving order information for other cargo storage and/or transportation services that may facilitate, and generating one or more task workflows based thereon. In embodiments, the workflow definition system uses output from task definitions, task parsing systems, and real-time external data, such as maintenance management systems, ERP systems, etc., to determine task workflows to identify workflow possibilities. Integrates information from smart container fleet management systems. In embodiments, a task workflow defines the order and manner in which tasks are performed for a project/task. In embodiments, a workflow definition system may apply job descriptive information to a set of task definitions and fleet configuration data to create one or more workflows for performing one or more activities of a job. As an example, a workflow could cover activities such as last-mile 3D printing of sneakers in a container. The tasks defined for this activity can be collected into a workflow or part thereof, sequenced to appropriately comply with the task requirements, and issued as a set of requirements for performing the activity/workflow. The task workflow definition defines the quantity and type of smart containers, 3D printers, robots, tools and/or end effectors, etc., that can be provided by the fleet configuration system 13020 for one or more tasks ordered by the workflow definition system. May include explanatory information. In embodiments, this portion of the workflow definition may be used to identify and determine required configuration, e.g., one or more smart containers to be primed prior to performing tasks in the workflow (e.g., if a smart container is to be deployed in the workflow). used by other modules of the task configuration system 13018 (e.g., order execution system 13022) to ensure that the module is (re)configured into a configuration capable of performing the task prior to performing the task defined in It can be. Other information generated from the order execution plan may include task sequences (e.g., generated by a workflow system), which may further identify the sequence of smart containers required to perform the task.

The workflow definition system can leverage the resources of the smart container configuration library when defining workflows. Workflow definition parameters, such as how to determine minimum time between tasks, coordination between tasks, task classification, workflow scope, etc., may be available in library 13044 and/or in information retrieved from freight storage and/or transportation service orders. You can. These and other parameters can be set to default values, but may also include task-specific variables that can be adjusted, for example, by a workflow definition system to meet task-specific requirements. An example of the use of smart container configuration library information to develop task workflow definitions includes a cargo unloading task (e.g., by a set of robotic arms attached to a smart container or by robots embedded within the container), and then may include self-stacked storage tasks (e.g., by on-container rail and/or lift systems). Useful information that a workflow definition system can utilize from the smart container configuration library may include templates, pre-configured or default workflows, such as workflows developed for previous execution of the task. The workflow definition system determines which workflows, if any, within library 13044 (default workflow) are suitable for use in the current job workflow definition instance; determine adjustments to the retrieved workflow; You can create instance-specific job workflows that can include additional tasks not found in the default workflow, and/or exclude unnecessary tasks found in the default workflow, etc.

Other examples of smart container configuration library information that may be useful in developing task workflow definitions include the availability of sensor detection packages. These sensor detection packages can indicate a preferred sequence of sensing tasks and thus influence the workflows of these tasks. These and related reconfigured sensor and detection packages can combine sensor selection, sensing, information collection, preprocessing, routing, integration, processing, etc. These sensor and detection packages may be included in the fleet configuration process, such as being included in an order execution plan for use by order execution, monitoring, and reporting system 13022. In embodiments, its use is indicated to provide various monitoring activities, etc.

The job workflow definition system is designed to, among other things, identify potential workflow independence and dependencies for constructing order execution plans that may involve parallelized use of fleet resources such as teams, task-to-task dependencies (e.g. For example, performing the second task of unloading cargo may be dependent on completing the first task of transporting the cargo to the destination container terminal.

Features of the intelligent service, such as digital twin capabilities, can also be advantageously applied to simulate and verify workflows, for example using the workflow simulation system of the work configuration system 13018. A workflow simulation system may perform simulations of portions of a work configuration, such as portions organized into work workflows by a workflow definition system. In an example of a workflow simulation, a set of tasks defined by a task definition system and organized as part of a task workflow may include a smart container digital twin (13504), a container ship digital twin, a shipping container port digital twin, a rail digital twin, and a truck digital twin. , can be modeled using functional equivalents for smart containers, tasks, workflows, etc., such as environment digital twin, task digital twin, workflow digital twin, team digital twin, fleet digital twin, etc. These digital twins can be retrieved from library 13044 and executed by a processor to simulate a set of tasks, such as verifying defined tasks. In embodiments, the fleet intelligence system provides at least a portion of such workflow simulation, such as by applying workflow definitions and task definitions to one or more workflow models and/or digital twins operating in an artificial intelligence environment machine learning environment. It can be used to do this.

A workflow simulation system may also generate feedback from simulating workflows defined by the workflow definition system, which may be useful for improving workflow definition, task definition, smart container selection, etc.

A workflow simulation system can establish or otherwise access criteria for determining whether a workflow meets criteria, such as successfully completing tasks, tasks, etc. in a timely manner. By applying these criteria for measuring the results of the workflow simulation, the workflow simulation system can determine one or more workflow options, passed to the workflow definition system, before providing feedback to, for example, the task definition system, task parsing system, etc. Smart container options, fleet configuration options, etc. can be verified. Options that do not meet the criteria (e.g. consume excessive resources, cause wear and tear on the smart container, fail to meet schedule, result in a high percentage of damaged cargo, etc.) can be implemented by adding tasks to workflows, etc. It can be marked to improve task organization functions such as structuring.

Additionally, workflow simulation systems can utilize system intelligence services. In embodiments, a system intelligence service may provide access to and operation of instances of fleet digital twin modules that may provide critical understanding of fleet-based impacts on workflow definition for performing requested tasks. In an embodiment, the logistics digital twin of the fleet intelligence system can provide useful workflow simulation information through modeling operations of delivery and costs of smart containers, personnel, and support equipment for smart container fleet services. This modeling of fleet logistics shows that a first local fleet that will soon become available (perhaps after the preferred start date of the requested work) can complete the work at a lower cost than using a currently available second fleet. It can be expressed. In embodiments, a fleet digital twin may facilitate identifying smart container operational assets available during scheduled operations by modeling fleet behavior, such as smart container maintenance requirements for smart containers during preferred order execution times. In embodiments, the fleet intelligence system's task digital twin capabilities may facilitate modeling of smart container cargo configurations, such as when smart container cargo is reconfigured during operations. The task digital twin capability of fleet intelligence systems further benefits workflow definition clarity through workflow simulation by applying a virtual set 13504 of pre-configured smart container digital twins to perform optionally defined candidate workflows or parts thereof. can give. In embodiments, the team digital twin capabilities of the fleet intelligence system can be implemented by the work organization system 13018, for example, by using pre-configured smart container teams to operate and validate candidate workflows prepared by the workflow definition system. It can benefit workflow simulation systems.

In embodiments, the results of a workflow simulation may include one or more data structures suitable for use in order execution planning.

In addition to task definitions, smart container definitions, workflow definitions, fleet configuration parameters, etc., the order execution plan may be structured/organized into contracts for the work, e.g., by or in association with the work composition system 13018. It can identify smart contracts, delivery delivery times to work resources (e.g., fleets of smart containers), schedules of deliverables, etc.

In an embodiment, fleet configuration system 13020 configures the fleet's resources for work based on task definitions, workflow definitions, etc. Fleet configuration system 13020 may consider cost, transportation mode(s) (e.g., container ship, rail, truck, container hyperloop, and/or self-driving container), environmental conditions, time constraints, smart containers, and/or components. Fleet composition can be determined based on other considerations, such as available inventory. Fleet organization system 13020 may operate in cooperation with work organization system 13018, such as when tasks are organized into workflows. Because job workflows may be affected by the availability of each type of smart container, job configuration system 13018 may utilize fleet configuration system 13020 when determining candidate job workflows.

In embodiments, configuring a fleet for requested work may include configuring fleet resources into smart container teams assigned to specific tasks and/or projects (smart containers or smart container teams may be comprised of multiple tasks and/or projects). Note that you can be assigned). Each smart container team may include one or more smart container operating units, which may include any one or more of smart containers, robots, humans, transportation modes, machines (e.g., 3D printers), tools, etc. there is. Additionally, the configured smart container team may be task-specific, and team membership may be temporary for any given smart container operating unit. By way of example, a robot configured to perform cargo loading and/or unloading operations may be assigned to a first smart container team only for the duration of time that cargo loading and/or unloading tasks are being performed by the first smart container team. The same robot may also be assigned to a second smart container team only for a duration of time during which the second smart container team cargo unloading and/or loading is being performed. Time sharing of fleet resources such as robots, container ships, trucks, rail, reach stackers, forklifts, cranes, etc. can be communicated, for example, from the fleet configuration system 13020 to the delivery configuration system, The workflow defined by the delivery configuration system can take into account the availability of cargo loading/unloading robots for each smart container team. In embodiments, any given smart container or group of smart containers can be distributed across multiple tasks by the fleet configuration system 13020 using a smart container-specific time-sharing approach or other resource utilization optimization technique. Can be assigned to teams. In one example, the fleet configuration system 13020 uses a multidimensional smart container utilization planning system to assign each smart container within a fleet to a task for a time unit, such as a day, hour, or portion thereof, of the shipment configuration system. Each instance can request the use of a smart container for a specific period of time. The fleet configuration system 13020 may respond to the request with a smart container fleet configuration description notifying task workflow definitions, etc.

In some embodiments, fleet configuration for the requested task may include smart containers such as edge devices, charging capabilities and/or charging stations, local data storage capabilities, container storage facilities, spare parts, human technicians, etc. May include allocating support resources.

In embodiments, fleet configuration system 13020 may utilize a library to determine fleet configuration. In these embodiments, the fleet configuration system 13020 may determine team configurations for defined tasks or projects using a library 13044 that defines different configurations to perform specific tasks, thereby looking up A table or other association is used to determine team compositions for a given set of tasks. In embodiments, library 13044 may include properties of different smart container types. As an example, the properties of a smart container may indicate the size or volume of the smart container. In embodiments, fleet configuration system 13020 may be configured to use smart smart devices that can perform tasks based on attributes identified by the task parsing system (and optionally configured as task definitions) and one or more cargo storage and/or transportation service order parameters. You can filter the type of container. When a task or work operation requires access through a tunnel smaller than the available smart containers (e.g., based on data generated by a task parsing system, existing order execution plans, freight storage and/or transportation service orders, etc.) , the fleet configuration system 13020 will not include smart containers; Instead, we will attempt to identify different smart containers and/or smart container types/configurations that can meet the tunnel size requirements. In embodiments, fleet configuration system 13020 may reference combinations of smart container sizes and/or types, etc. to fit the requirements of a defined task. Additionally, the fleet configuration system 13020 may propose two smart containers to perform a task if other requirements of the task may not be met. In embodiments, fleet configuration system 13020 may pass a fleet definition including a plurality of smart containers, smart container types, smart container configurations, etc. to task configuration system 13018. A general goal of the fleet configuration system 13020 may include creating a fleet configuration(s) that requires the fewest smart containers and/or smart container types for proper execution of a portion of an order. However, fleet configuration system 13020 may operate in conjunction with a task definition system to create a task-specific fleet configuration that includes more than one smart container type/configuration combination, thereby Allows other elements to efficiently manage the execution of requested tasks. This fleet configuration includes smart containers, etc., that other elements of the delivery configuration system (e.g., a task workflow creation system) can take into account, for example, when organizing a plurality of defined smart container tasks into task workflows. It may indicate desirable smart containers and/or smart container combinations to meet goals such as efficient use. Accordingly, a fleet configuration may include marking primary, secondary, and tertiary smart containers to perform tasks. Alternatively, a fleet configuration for ordering cargo storage and/or transportation services may identify multiple smart containers, each assigned a utilization weight based on criteria such as efficient task completion, profitability, fleet smart container usage preference, etc. do.

In an embodiment, fleet configuration system 13020 consults an inventory data repository to determine available smart containers and/or modules (e.g., physical modules and/or software modules) for configuring smart containers, and You can configure the location of the container and/or part, the condition of the part (e.g., whether maintenance is expired or required for an available smart container or part), etc. In this way, fleet composition for jobs, tasks, teams, etc. can be determined by available inventory of smart containers, modules, support equipment and/or spare parts. Additionally, the fleet maintenance management system described herein may provide aspects of smart container status that may be added and/or replenished to the inventory data store, such as which smart containers are being reserved from use for critical maintenance. , some smart containers may be deployed but have reduced capacity due to service and/or maintenance or other concerns, status of spare parts, or other service activities (e.g. due date, current location, expected installation, etc.) You can track whether you can have it. Accordingly, the fleet configuration system 13020 may be consulted and/or notified by the fleet maintenance system for fleet resource maintenance knowledge that may impact operations. Additionally or alternatively, fleet configuration system 13020 may request fleet configuration from intelligence service 13004, where artificial intelligence modules 13404 include task definitions, workflow definitions, budget, and environment definition as input. , a task timeline, etc., may receive a set of parameters, evaluate a plurality of candidate flit configurations, and determine a target flit configuration capable of performing the task. In embodiments, humans may define or redefine any part of the fleet configuration through the human interface of the fleet configuration system.

In embodiments, job and fleet configurations may be fed to a digital twin system, whereby the digital twin system can perform a simulation of the job given the job and fleet configuration. Job configuration system 13018 and/or fleet configuration system 13020 optimizes (or substantially optimizes) one or more parameters, such as job timeline, overall cost, smart container downtime, maintenance-related downtime, transportation cost, etc. To achieve this, task configuration and fleet configuration can be redefined iteratively. Once the job configuration system 13018 and the fleet configuration system 13020 determine the task and workflow definitions as well as the fleet configuration, the smart container fleet management system can output an order execution plan corresponding to the cargo storage and/or transportation service orders. You can.

In embodiments, fleet configuration system 13020 may utilize digital twins when configuring fleet resources. Use of digital twins with a fleet configuration may include identifying and/or defining one or more digital twins of one or more smart containers based on information in a task definition. Fleet configuration may include identifying the configuration and/or operation of the smart container so that the smart container can perform the root and/or task or portion thereof. These smart container task configuration instructions may be generated through the use of a digital twin for one or more of the set of candidate smart containers to perform the task. In an illustrative example, a smart container may be associated with a plurality of configuration/action data structures to configure the smart container to perform routines, actions, roots, tasks, etc. Fleet configuration system 13020 may identify or otherwise be provided with one or more candidate smart container configuration data structures (e.g., from library 13044) for use in performing the task. Part of this candidate configuration data structure may include travel speeds for moving up a ramp onto a container ship. Requested task requirements may explicitly or implicitly indicate that the movement speed is different from the value in the candidate configuration data structure. In an embodiment, the fleet configuration system makes arbitrary adjustments to the candidate configuration data structure (e.g., reduces the movement speed) and applies them to an instantiation of the digital twin of the candidate smart container with the adjusted configuration data structure. The execution (e.g., simulation) of the digital twin may be observed and/or evaluated, and stored in a library 13044, etc. Newly stored configuration data structures may be categorized based on other parameters, such as cargo storage and/or transportation service orders and/or requested operations, tasks, etc., to achieve efficient access in the future.

The Smart Container Configuration Library provides job information, smart container information, fleet information, task definition rules/metadata that can be useful in determining how to define smart container tasks, workflow configuration rules and/or techniques, and shipping configuration system information. Results of previous cargo storage and/or transportation service orders from the application (e.g., previous order execution plans), etc. Library 13044 may be accessed and/or updated by functions of the operating system. Illustrative examples of library 13044 are variously described herein with respect to job operating system functions and features, such as job configuration. By way of example, a smart container configuration library may contain specific references to configurations of smart containers that can be used during fleet configuration, order execution, etc. In this example, the smart container configuration library will have a reference to the smart container configuration data set (e.g., data that, when uploaded to a smart container, allows the smart container to perform functions such as 3D printing, in-container packaging, etc.) You can. Additionally, the library may provide cross-references of smart container configurations with other smart container-related information, such as base model, version, required features, etc., that may be required for successful deployment of a smart container configured with a given configuration. Additionally, the library may suggest alternatives for specific combinations of smart containers and configurations, such as indicating that newer versions of the smart container model may include built-in capabilities provided by specific configurations. Accordingly, the fleet composition system can have greater flexibility in deciding which smart containers to deploy for different tasks. References to library 13044 are made herein using context modifiers, such as smart container configuration libraries. These context recipients may suggest one or more portions and/or instances of library 13044 for illustrative purposes only.

In embodiments, optimization features of the task definition system are described below with respect to feedback from other elements, such as the task composition system 13018, such as a workflow simulation system.

In some embodiments, the fleet operating system and fleet intelligence system perform feedback for order execution time iterations of component activities, such as to adapt and execute instances of an order execution plan. In these embodiments, feedback within work configuration system 13018 facilitates repetition of configuration activities when creating components of an order execution plan, such as task definitions and workflow definitions. As described for these embodiments, intelligence services 13004 may be used for at least these iterations. However, it is contemplated that the resources of intelligence service 13004 may also, or additionally, be used to enhance execution of the order execution plan.

In embodiments, order execution system 13022 of fleet management system 13002 may receive an order execution plan from work configuration system 13018, for example, in response to a freight storage and/or transportation service order. Order execution system 13022 advances the plan, activates and monitors smart container units and other fleet resources, and provides feedback, optionally real-time feedback, for example, based on smart container unit monitoring data. It can make performance easier. This feedback may be processed by the artificial intelligence capabilities of intelligence service 13004 to determine adjustments to the order execution plan, for example, task definitions, etc. When feedback and adjustments are performed in real time or near real time (e.g., in advance of upcoming order execution activities, such as steps in a workflow), the functionality of work organization system 13018 is dependent on the current order execution system 13022. It may be iterated to modify an existing order execution plan, such as an instance of the plan being executed.

The artificial intelligence systems of intelligence service 13004 may perform simulations and use the results of the simulations as input to task composition system 13018 to update corresponding task definitions. In embodiments, the fleet intelligence system alerts the fleet management system about the need to adapt such task definitions, which can be used by the system to update pre-configured task definitions stored, for example, in the smart container task library 13044, etc. It can be sent to (13002). These alerts can be used by the fleet operating system to coordinate with the order execution system 13022 so that pending tasks are not executed before being refreshed in the order execution plan. In embodiments, work organization system 13018 may release only portions of the order execution plan to order execution system 13022 so that unreleased portions can be adapted, thereby allowing updates to the execution plan. Mitigate the impact on the order execution system, such as requiring work to be halted, delayed, or otherwise weakened while it is taking place.

Examples of task organization, etc., presented herein generally consider a single task to be composed by the task organization system 13018, but many tasks may be composed simultaneously. Methods and systems for real-time or near-real-time feedback described herein address the task isolation requirements necessary to support simultaneous processing of cargo storage and/or transportation service orders from different entities (e.g., The identification data may be obfuscated), while feedback on the task definition of the first task may be applied to any instance of the task configuration activity being performed such that it may be beneficial to the task definition of the second task.

In embodiments, capturing data indicative of completion of a requested task may include extracting such data from a task completion data set. This work done data set can be organized to facilitate identifying information that may be useful for learning and optimization. In one example, a work-complete data set may be designated, such as by the use of metadata tags, logical and/or physical separation, or other indicative data that indicates exceptions or major variations from a prediction. In an example, upon completion of a task, the percentage of damaged packages returned by the smart container may exceed an expected and/or acceptable number. This excessive count of corrupted packages can be extracted and flagged as candidate information for learning and optimization feedback to be sent to intelligence service 13004. In embodiments, an order execution plan may consist of an indicator of the type of data that will be collected and used for learning and optimization feedback. Intelligence service 13004 may recommend to work configuration system 13018 the type of data to be so presented based on other factors known to the fleet intelligence system, such as queries made by the smart container design engineering team, etc. In embodiments, learning and optimization feedback is provided to the fleet intelligence service to perform optimization of the artificial intelligence service (e.g., recommending smart container teams, smart container types, workflows, etc.), among other things. It can be used by . Referring to the description herein, pre-configured tasks, smart container configurations, team configurations, etc. may be retrieved from the library 13044. As these preconfigured aspects of the order execution plan are executed, data representative of performance may be flagged for use as learning and optimization feedback to continuously improve these preconfigured aspects. The results of using this data include preconfigured tasks adapted to field conditions that can perform better in the real world. Other uses of this data include improved digital twins and machine learning models.

In embodiments, the job description to be parsed may include relevant job description details, goals, objectives, requirements, preferences, etc., and may be described elsewhere herein. Not all relevant task information will be included within the request, but one or more links to auxiliary task description data may be included. Ancillary task data may be stored remotely (e.g., accessed via an Internet URL in the task description) from the freight storage and/or transportation service order data set. Optionally, auxiliary task data may be stored in a data structure accessible to smart container system 13000, such as a fleet library 13044, requestor-specific storage, etc. Secondary task data includes formal standards (e.g., local nuisance regulations, safety (OSHA), electrical (NEC), quality, etc.), acceptance requirements (e.g., form, steps, timing, dependency on other tasks, etc.); Details of the legal requirements of the operation (e.g. customs requirements, relevant laws, etc.), shipper standards (e.g. acceptable percentage of damaged cargo), industry norms (e.g. work times, material selection, templates, etc.) , approved vendors (e.g., from whom supplies and other consumables will be obtained), references to pre-configured tasks, user interface templates/menus/screens for each aspect of the task (e.g., user requests for status) methods, observing activities, changing task requirements, responding to inquiries, etc.). Freight storage and/or transportation service order data 13096 and, if indicated, auxiliary task data are processed by a task definition collection facility that operates in conjunction with the task parsing system to generate task instance specific content. Such task instance-specific content may include, among other things, initial sequence timings that may be defined in the input data (e.g., "perform task A before task B") and/or initial sequence timings that may be derived therefrom. (For example, securing cargo in a container with steel strapping must be done after the cargo has been loaded into the container). The job parsing system may interact with the data processing system 13024 in transforming job description data to utilize information derived from smart container fleet management system accessible libraries, such as job and fleet libraries 13044. The collection facility may store some task description content, such as task identification information, links to internal auxiliary data, etc., directly into task instance storage.

In embodiments, one or more human interaction capabilities to facilitate task parsing and task definition may be used to collect information for pre-formatting, organizing, and inspecting tasks and task data (e.g., text entry, dialogue, etc.). May include knowledge-based systems (e.g., AI-based, etc.) that can interact with humans (via bots, haptic input, etc.). These interactions can replace or supplement receiving job descriptions.

The task parsing system may use task description information generated by or passed through the collection facility to construct task instance content appropriate for the task definition. The operational parsing system may use information provided by the collection facility to query content within library 13044 (e.g., optionally via data processing facility 13024 as shown). Content in the library that can be useful or informative in task definition includes job syntax (e.g., given job, job type, set of tasks, smart container types, smart container capabilities (e.g., type , size, cost, availability, etc.), keyword-to-task cross-references, workflow definition rules, terms related to order execution plan format/content/structure). Additionally, the library provides example smart container configurations (e.g., based on task keywords, etc.), example team configurations (e.g., to perform specific types or classes of tasks), task definitions, workflows, and workflow definitions. , may include templates for various task definition-related activities, such as exemplary order execution plan(s).

The keyword-based task lookup module retrieves information in the job instance repository, such as task-oriented keywords, and applies them to the library 13044 to create preconfigured or templated tasks or parts thereof. can potentially be identified. As an example, a job description may include keywords such as “submerged,” which may suggest a need for smart containers configured to move underwater. When these keywords are combined with the action “underwater 3D printing,” the keyword-based task lookup module can identify smart container types that can perform 3D printing and move underwater. If the descriptor of a task in the library aligns with one or more task description keywords, the task may be considered a candidate task for the task.

In embodiments, the task definer module may process information in the job instance storage as well as candidate tasks provided by the task lookup module to form definitions for tasks to be performed by one or more smart containers. Defining a task may include tasks predefined by standards, laws (eg, tariffs), etc. Each task definition can contain information useful for identifying the smart container type to perform the task.

In embodiments, the task definition system may store cargo in the context of a smart container type (e.g., as provided by a cargo storage and/or transportation service order parser) by identifying characteristics of the smart container type that align with the task data. and/or process task data derived from transportation service orders. In an example embodiment, the task definition system may determine that the task data represents characteristics of a marine smart container that may include movement in an arctic environment. In this example, the task definition system may create a task definition for the delivery task that includes at least the requirements for smart container selection based on these characteristics. In these example embodiments, the task definition may further include a desired degree of tolerance for cold temperatures. The task definition system may also define characteristics of one or more smart containers (e.g., based on task information derived from cargo storage and/or transportation service orders) that may not be suitable for integration into a single smart container/smart container type. You can decide. This decision may be based, for example, on smart container properties and type data accessible in library 13044. In this example, the task definition system may define multiple tasks, each of which has smart container properties that match smart container property information in library 13044. In an embodiment, the task definition system may optionally configure one of the tasks to require each type of a number of incompatible smart container features that the fleet configuration system 13020 may use when configuring fleet resources, such as smart containers, etc. With the above indications, it is possible to define tasks with a number of potentially incompatible smart container characteristics. In embodiments, the task definition may be, for example, an arrangement of task requirements (e.g., derived from task information of a cargo storage and/or transportation service order), smart container properties, and use in library 13044. Based on the possible smart container types, it may include one or more suggestions for types of smart containers to perform the task. As described below, fleet composition system 13020 may evaluate a task definition that includes any proposed smart container types. Other example data that may be passed when defining a task may include task sequence dependencies that may be suitable for defining a workflow that includes the defined task. As an example, a container self-cleaning task may be required to be performed after an unloading operation. These dependencies are documented and can be relied upon by the workflow definition system. The task definer module may store the defined task in job instance storage, where the defined task may be cross-referenced against job description data (e.g., keywords, etc.) such that future detections of the cross-referenced keywords may be It can quickly result in appropriate task definitions.

In an embodiment, fleet configuration system 13020 provides specific software, hardware, and smart container configuration requirements for completion of an order execution plan. In this example configuration, the fleet configuration proxy module may be configured to receive task definitions from task configuration system 13018. A fleet configuration proxy module may be instantiated in connection with the processing of cargo storage and/or transportation service orders by the operations configuration system 13018 to facilitate access to and use of fleet configuration system 13020 resources and systems. there is. These and other instantiations of the fleet configuration proxy module are further described herein with respect to the task configuration system 13018. The fleet configuration proxy module may process task definitions and forward them to fleet resource identification systems, such as the fleet smart container operating unit identification and fleet non-smart container operating unit identification systems. Each of these identification systems can process task definition data provided through a fleet configuration proxy to separate operational data from fleet resource data. The task definition describes the set of fleet resources required to perform the task, such as types of smart container operating units, support resources (e.g., power systems, robots, cranes, communication systems, etc.) can do. The smart container operating unit type identification system may provide task-specific smart container operating unit requirement data to the fleet configuration scheduler. Task-specific smart container operating unit demand data may identify the type and quantity of smart containers, the specific smart container operating unit (e.g., by unique identifier), smart container operating unit capabilities, etc.

In some embodiments, a fleet configuration scheduler may respond to cargo storage and/or transportation service orders by allocating fleet resources to meet cargo storage and/or transportation service order demands. These requests may be preprocessed by the task configuration system 13018, as described herein, and specifically by the task definition system to facilitate fleet configuration, allocation, and scheduling. The fleet configuration scheduler processes inputs describing fleet inventory, such as smart container operating unit inventory and traditional container operating unit inventory, to identify candidate inventory elements to fulfill cargo storage and/or transportation service orders. These inventories can be adjusted based on existing allocations of smart container operating units and traditional container operating units. As an example, all smart containers of a type identified in the smart container operating unit task-specific demand data may be assigned over a constrained time duration for the requested tasks to be performed. The fleet configuration scheduler (e.g., with assistance from other system resources such as intelligence services 13004, resource provisioning system 13014, etc.) Based on the equivalent data available to the fleet configuration scheduler, it can assign smart containers for the requested activities. To accomplish this assignment, intelligence service 13004 is provided with information describing the functionality to be provided by the smart container indicated in the task-specific demand data and information describing the tasks and/or activities to be performed by the smart container. It can be. Other context may also be available to the intelligence service 13004, such as differences in specifications for performing tasks by appropriately configured smart containers. Through the use of artificial intelligence, which may include determining the impact on overall cargo storage and/or transportation service orders based on the use of two different smart container types, intelligence service 13004 provides smart container replacement guidance to fleets. Can be provided to the configuration scheduler. Such guidance may result in the allocation of smart containers and necessary configuration data/characteristics for use when executing order execution plans corresponding to cargo storage and/or transportation service orders that prompted these fleet configuration scheduling activities.

In embodiments, the task definition may include recommendations for one or more types of smart containers (e.g., based on task requirements, smart container characteristics, and sorting of smart container types), and may include recommendations for preferred The type can be specified in the task definition.

The fleet configuration scheduler may rely on other fleet systems, such as the resource provisioning module 13014, which may contribute to the provisioning of fleet and third-party resources and supplies and/or determine provisioning.

Other fleet systems, including the intelligent service 13004, resource provisioning module 13014, and the fleet configuration scheduler, provide a list of fleet configuration options that may be considered by the fleet configuration scheduler when configuring the fleet in response to task configuration activities, etc. It can interact with a fleet configuration modeling system that can facilitate its creation. The fleet configuration modeling system may provide simulation of fleet configurations, such as by using fleet digital twins, which may optionally be associated with digital twin module 13420 of intelligence service 13004.

In embodiments, the fleet configuration scheduler may rely on a fleet team organizer module to help determine and/or achieve team configurations. Task-specific data may identify (e.g., recommend) a set(s) of smart container operating units to be organized into teams. Additionally, task-specific data may represent information that may indicate teaming, such as co-location of smart containers at a container terminal. The Fleet Team Organizer module can identify and/or specify team metadata to use when organizing a fleet. Team metadata may indicate team membership and a time frame for membership (e.g., from one date to another date, from the start of a task until the task is completed, etc.).

The Fleet Configuration Scheduler configures fleet allocation data, such as fleets in the Fleet Smart Container Operating Unit Allocation Dataset and the Non-Container Operating Unit Allocation Dataset, with Fleet Configuration Allocation Information based on the configuration(s) generated for the provided task-specific demand data. The set (which may be used by the fleet resource allocation and/or reservation capabilities described herein) may be updated. Various inputs, including fleet configuration (e.g., weather, location data, traffic data, industry standards, task-specific context information, etc.) that affect external data 13036, are sent to, among other things, the Fleet Configuration Proxy module. may be optionally processed repeatedly by the fleet configuration scheduler to generate fleet configurations that can be returned to a running instance of the task configuration system 13018.

In embodiments, a workflow definition system may be configured to create a definition of a workflow for a requested task using the resources of a smart container fleet management system. The configuration of the workflow definition system includes task definitions, which may be provided from a task definition system or sourced from a library 13044, and interacting with the fleet configuration system 13020 (e.g., via the fleet configuration proxy module). ) may include a collection module that receives and processes task-specific fleet configuration information that may be provided from the task configuration system 13018.

Collecting task definitions and/or fleet configuration information may include aligning the fleet configuration information with one or more task definitions. As an example of aligning tasks with flit configuration information, flit configuration information may be tagged as applying to one or more tasks within a set of collected task definitions, such as using an identifier of the task or tasks. Other ways to align task definition(s) with fleet configuration information include, for example, when a fleet configuration reference/value is received simultaneously with a task definition, so that the collection module can mark these two data items as aligned. This may be based on the timing of collection. Other ways to align task definition(s) with fleet configuration information may include one or more data values within the task definition, such as a data set, linked list, flat, etc., representing the fleet configuration information by which the task(s) should be aligned. It can be a flat file, a structured data set, etc. The fleet configuration information may include one or more task identifiers to which the fleet configuration information is associated and/or should be applied when creating a workflow definition.

Collection may further include processing references to workflow content within library 13044 (e.g., URLs, hyperlinks, external names, etc.) that may be found in any collected content. . In one example, the task definition may include the name of the task stored in library 13044. The collection module can configure its syntax (e.g., a prefix may be added to the task identifier indicating that the task will be retrieved from a library) and/or structure the task definition (e.g., a subset of tasks may be retrieved from the library). The name may be identified by a list of task names stored within a subset of the task definition structured to indicate that it is to be searched. Examples of aggregation herein relate to instances of aggregation of one or more task definitions, but aggregation may be performed on batches of tasks. Multiple instances of a collection module may be instantiated and operate simultaneously to process multiple task definitions that may be performed. Optionally, a stream of task definitions can be received by collection, and each task in the stream is collected sequentially.

One or more results of processing by the collection module may include a task dependency determination module that can determine dependencies between tasks, such as which tasks need to be performed in sequence and which tasks can be performed independently of other tasks. A set of workflow definition activities may be provided, including: The task dependency determination module may also determine the dependency of tasks on other factors such as availability of fleet resources, calendar/date/time, readiness of supply materials, etc. Dependencies on other arguments can be identified in the task definition, such as by marking a given task state as the starting point for the task. Additionally, dependencies on other factors may be due to the task definition given during collection (e.g., based on aligning tasks with the fleet configuration, establishing dependencies on the availability of fleet resources such as special-purpose smart containers, etc.). there is.

A task grouping activity can process the results of task dependent activities to create groups of tasks based on a range of criteria, such as tasks that depend on a given task being completed can be grouped for concurrent execution. Grouping tasks can be based on dependency on fleet resource availability, such that tasks dependent on a fleet resource can be grouped and performed when the resource is available. The execution order of these grouped tasks may be based on dependencies between tasks. Typically, tasks are evaluated based on factors such as cost reduction, resource guarding, task priority, available order execution funds, expected fleet resource maintenance needs, earliest task start/finish time, latest task start/finish time, etc. Can be grouped for various purposes.

The Task Workflow Step Definition activity may determine which task(s) can be organized into each step of one or more workflows. Based on inter-task dependencies (or lack thereof), multiple workflows may be defined, each workflow containing one or more workflow steps defined in the Define Workflow Steps activity. Additionally, a workflow step, once defined, can be assigned and/or referenced to multiple workflows. When dependencies exist, such as the availability of smart containers to perform tasks in a workflow step, multiple workflows themselves can become dependent. In this workflow, the performance of other tasks can be performed simultaneously even if the initial task of opening a port must be performed sequentially due to fleet resource utilization dependencies.

In embodiments, the defined workflow step may be an adapted variant of a candidate workflow step, such as a workflow step retrieved from library 13044. The workflow step definition activity may be performed by other fleet resources, such as data processing system 13024 and/or artificial intelligence module 13404, to adapt candidate workflow steps for use in defining one or more workflow steps for a given task. May request input from system services.

Information such as workflow step dependencies may be utilized by the workflow step link activity, which may receive step link recommendation(s) from intelligence service 13004, etc. A workflow step link activity may create a data structure representing a sequence of performing defined workflow steps (e.g., a workflow definition). A workflow definition captures task-specific workflow information such as workflow step sequencing, workflow step performance sequence, workflow step independence, step-by-step links to workflow steps, workflow success criteria, cross-workflow dependencies, etc. Can contain data.

In embodiments, the workflow definition(s) may be stored in job instance storage, where the workflow definition(s) may be referenced as needed during job configuration and/or order execution. The workflow definition may be stored in a fleet library 13044, where the workflow definition may be referenced by other operations by third parties, such as shippers. Workflow definitions may be stored elsewhere (e.g., a cloud storage facility) based on architectural considerations, such as being distributed to edge computing infrastructure resources proximate to work deployment sites, etc.

In embodiments, workflows may be simulated as indicated in the description of task configuration system 13018. The results of the simulation can be directed to an aggregation module where aggregation operations, such as sorting task description data and fleet configuration data, can be improved. Results can also be passed on as feedback to other components of the system to improve task definition, job composition, fleet composition, etc.

Figure 147 shows a digital twin module according to some embodiments of the present invention. In an embodiment, the digital twin module 13420 may be configured to create a set of shipping digital twins 13302 (e.g., an individual smart container digital twin 13504, a smart container fleet digital twin, a container ship digital twin, a container port digital twin, a smart Create a container fleet manager or operator digital twin, shipper digital twin, etc.) In an embodiment, the digital twin module 13420 may use sensor data obtained from each of the sensor systems 13304 that monitor the delivery digital twins 13302, for example, to create individual delivery digital twins ( 13302) maintains a set of states. In embodiments, digital twin module 13420 may include digital twin management system 13306, digital twin I/O system 13308, digital twin simulation system 13310, digital twin dynamic model system 13312, and/or digital twin It may include a control module 13314. In embodiments, digital twin module 13420 may provide a real-time sensor API that provides a set of capabilities to enable a set of interfaces to the sensors of individual sensor systems 13304. In embodiments, digital twin module 13420 may use other suitable APIs, brokers, connectors, bridges, gateways, hubs to facilitate the transfer of data to and from digital twin module 13420. , ports, routers, switches, data integration systems, peer-to-peer systems, etc. In these embodiments, these connection components allow IIOT sensors or intermediary devices (e.g., relays, edge devices, switches, etc.) within sensor system 13304 to communicate data to digital twin module 13420 and/or It may be capable of receiving data (e.g., configuration data, control data, etc.) from the twin module 13420 or another external system. In an embodiment, digital twin module 13420 may further include a digital twin datastore 13316 that stores delivery digital twin 13302.

The digital twin may be an individual smart container 13026, a fleet of smart containers 13026, a shipping environment (e.g., a physical location along the route, a shipping container port or terminal, a smart container charging station, a shipping yard, a container storage facility, may refer to a digital representation of one or more shipping entities, such as a container ship, a smart container machine, a physical object, a device, a sensor, a human, or any combination thereof. Non-limiting examples of physical objects include smart container cargo, physical barriers along the route, raw materials, manufactured products, excavated materials, boxes, bins, cooling towers, vat, pallets, barrels, and pallets. ), bin, and more. Non-limiting examples of devices include robots, computers, vehicles (e.g., ships, automobiles, trucks, trains, etc.), machines/equipment (e.g., forklifts, cranes, reach stackers, packaging systems, sorting systems, tractors, including tillers, drills, presses, assembly lines, conveyor belts, etc.). Sensors may be any sensor devices and/or sensor aggregation devices found in a sensor system in the environment. Non-limiting examples of sensors that can be implemented in a sensor system include temperature sensors, humidity sensors, vibration sensors, LIDAR sensors, SLAM sensors, SONAR sensors, motion sensors, chemical sensors, audio sensors, May include pressure sensors, weight sensors, radiation sensors, video sensors, wearable devices, relays, edge devices, crosspoint switches, and/or any other suitable sensors. You can. Examples of different types of physical objects, devices, sensors, and environments are referenced throughout this disclosure.

In some embodiments, on-device sensor fusion and data storage in smart containers are supported, where data from multiple sensors is multiplexed at the device for storage of a fused data stream. For example, pressure and temperature data can be stored in a byte-like structure (where time, pressure and temperature are bytes in the data structure, and thus pressure and temperature remain linked in time, without requiring separate processing of the streams by external systems), or by combining pressure and temperature in time series, by addition, division, multiplication, subtraction, etc. Can be multiplexed into data streams. Any of the sensor data types described throughout this disclosure can be fused in this way and stored in a local data pool, in storage, or on an IoT device such as a data collector, component of a machine, etc.

In some embodiments, a set of digital twins may represent an entire organization, such as shipping lines, container terminal operators, shippers, manufacturers, energy production organizations, regulatory organizations, governments, etc. In these examples, the digital twins may include digital twins of one or more facilities of the organization.

In an embodiment, digital twin management system 13306 creates a digital twin. A digital twin can be composed of other digital twins (e.g., through references). In this way, a discrete digital twin can be composed of a set of other discrete digital twins. For example, the digital twin of a smart container may be the digital twins of sensors on and within the smart container, the digital twins of the components that make up the smart container, the digital twins of the smart container cargo, or the digital twins of the smart container cargo, integrated into or connected to the smart container. may include digital twins of other integrated devices (e.g., robots, 3D printers, packaging systems, or sorting systems), etc. Taking this example one step further, a digital twin of a container terminal is a layout of the terminal that includes the arrangement of physical assets and systems in or around the terminal, a digital twin of the shipping entities within the terminal, as well as a digital twin of the storage areas within the terminal. It may include a digital twin that represents it. In this second example, the container terminal's digital twin may reference embedded digital twins, which may then reference other digital twins embedded within these digital twins.

In some embodiments, a digital twin is comprised of workflows and/or processes, including the inputs, outputs, sequences of steps, decision points, processing loops, etc. that make up such workflows and processes. Can represent the same abstract entities. For example, a digital twin can be a digital representation of a logistics workflow, cargo loading process, cargo unloading process, customs process, 3D printing process, smart container energy charging process, etc. In these embodiments, the digital twin may include references to delivery entities included in the workflow or process. A digital twin of a delivery process can reflect the various stages of the process. In some of these embodiments, digital twin module 13420 receives real-time data from smart containers, shipping entities, and/or shipping environments where shipping processes occur (e.g., from sensor systems in a container port facility) and processes processes in real time. It may reflect the current (or substantially current) state of.

In embodiments, the digital representation is a data structure ( For example, it may contain a set of classes). For example, the set of properties of a smart container may include the type or class of the smart container, the dimensions of the smart container, the mass or weight of the smart container, the material(s) of the smart container, the physical properties of the smart container material(s), and the smart container's It may include the surface, the state of the smart container, the location of the smart container, identifiers of other digital twins contained within the smart container, and/or other suitable properties.

Examples of the behaviors of the smart container include the maximum acceleration of the smart container, the maximum speed of the smart container, the movement ranges of the smart container, the heating profile inside the smart container, the cooling profile inside the smart container, the processes performed by the smart container, the It may include operations performed by the container, etc.

The set of properties of a physical object includes: type of physical object, dimensions of the physical object, mass or weight of the physical object, density of the physical object, material(s) of the physical object, physical properties of the material(s), physical object It may include the surface of the physical object, the state of the physical object, the location of the physical object, identifiers of other digital twins contained within the physical object, and/or other suitable properties.

Examples of the behavior of a physical object include the state of the physical object (e.g., solid, liquid, or gas), the melting point of the physical object, the density of the physical object when in a liquid state, and the state of the physical object when in a liquid state. Viscosity, freezing point of a physical object, density of a physical object in a solid state, hardness of a physical object in a solid state, malleability of a physical object, buoyancy of a physical object, conductivity of a physical object, combustion point of a physical object , how humidity affects a physical object, how water or other liquids affect a physical object, the terminal velocity of a physical object, etc.

The set of properties of a device includes the type of device, the dimensions of the device, the mass or weight of the device, the density of the device, the material(s) of the device, the physical properties of the material(s), the surface of the device, the output of the device, the state of the device, It may include the location of the device, the trajectory of the device, the vibration characteristics of the device, identifiers of other digital twins to which the device is connected and/or contains, etc.

Examples of behaviors of the device include the maximum acceleration of the device, the maximum speed of the device, the ranges of movement of the device, the heating profile of the device, the cooling profile of the device, processes performed by the device, operations performed by the device, etc. can do.

Exemplary properties of the delivery environment include environmental dimensions, environmental boundaries, environmental temperature, environmental humidity, environmental air currents, physical objects within the environment, environmental currents (if a body of water), etc. can do. Examples of behaviors of the environment may include scientific laws that govern the environment, processes performed in the environment, rules or regulations that must be followed in the environment, etc.

In embodiments, properties of the digital twin may be adjusted. For example, the temperature of the digital twin, the humidity of the digital twin, the shape of the digital twin, the material of the digital twin, the dimensions of the digital twin, or any other suitable parameters can be adjusted. When properties of a digital twin are adjusted, other properties may also be affected. For example, as the temperature within the smart container increases, the pressure within the smart container, such as the pressure of a gas, may also increase according to the ideal gas law. In another example, if a digital twin of a sub-zero shipping environment is increased beyond freezing temperatures, the properties of the twin of buried water in a solid state (i.e., ice) may change over time to a liquid state.

A digital twin can be expressed in many different forms. In embodiments, a digital twin may be a visual digital twin rendered by a computing device so that a human user can interact with a fleet of smart containers, individual smart containers, physical objects (e.g., cargo, etc.), devices, sensors, and/or the delivery environment. You can see the digital representation of . In embodiments, the digital twin may be rendered and output to a display device. In some of these embodiments, the digital twin may be rendered in a graphical user interface (e.g., a scalable vector graphics (SVG) enabled user interface) to allow a user to interact with the digital twin. For example, a user can "drill down" on a specific element (e.g., a smart container) to view additional information about the element (e.g., the state of the smart container, the properties of the smart container, etc.). You can. In some embodiments, the digital twin may be rendered and output on a virtual reality display. For example, a user may view a 3D rendering of the delivery environment (e.g., using a monitor, augmented reality headset, or virtual reality headset). While doing so, users can observe/inspect digital twins of smart containers, physical assets, devices, etc. within the environment.

In some embodiments, the data structure of visual digital twins (i.e., digital twins configured to be displayed in 2D or 3D fashion) includes surfaces (e.g., splines, meshes, polygon meshes, etc.) may include. In some embodiments, surfaces may include texture data, shading information, and/or reflection data. In this way, the surface can be displayed more realistically. In some embodiments, these surfaces may be rendered by a visualization engine (not shown) when the digital twin is within view and/or when present in a larger digital twin (e.g., a digital twin in a shipping environment) . In these embodiments, digital twin module 13420 may render surfaces of digital objects, whereby the rendered digital twin may be depicted as a set of adjacent surfaces.

In embodiments, a user may provide input to control one or more properties of the digital twin through a graphical user interface. For example, a user can provide input that changes the properties of a digital twin. In response, digital twin module 13420 may calculate the impact of the changed attribute and update the digital twin and any other digital twins affected by the change in the attribute.

In embodiments, a user may view a process being performed on one or more digital twins (e.g., last-mile customization of a product via 3D printer within a smart container, cargo inspection, cargo classification, etc.). In these embodiments, the user can view the entire process or specific steps within the process.

In some embodiments, the shipping digital twin (and any digital twin embedded therein) may be represented in a non-visual representation (or “data representation”). In these embodiments, the digital twin and any embedded digital twins exist in a binary representation, but the relationships between the digital twins are maintained. For example, in embodiments, each digital twin and/or its components may be represented by a set of physical dimensions that define the shape of the digital twin (or its components). Additionally, the data structure that implements the digital twin may include the location of the digital twin. In some embodiments, the location of the digital twin may be provided as a set of coordinates. For example, a digital twin of a delivery environment may be defined in terms of a coordinate space (e.g., Cartesian coordinate space, polar coordinate space, etc.). In embodiments, embedded digital twins may be represented as a set of one or more ordered triples (e.g., [x coordinate, y coordinate, z coordinates] or other vector-based representations). In some of these embodiments, each aligned triple represents a specific point (e.g., center point, top, etc.) on the shipping entity (e.g., smart container, object, device, sensor, etc.) with respect to the environment in which the shipping entity resides. point, bottom point, etc.) can be displayed. In some embodiments, the data structure of the digital twin may include vectors representing the movement of the digital twin with respect to the environment. For example, a fluid (e.g., a liquid or gas) or a solid may be represented by a vector representing the velocity (e.g., the direction and magnitude of the velocity) of the entity represented by the digital twin. In embodiments, vectors within a digital twin may represent tiny subcomponents, such as particles in a fluid, and the digital twin may represent physical properties such as displacement, velocity, acceleration, momentum, kinetic energy, vibration properties, thermal properties, electromagnetic properties, etc. You can.

In some embodiments, a set of two or more digital twins may be represented by a graph database containing nodes and edges connecting the nodes. In some implementations, an edge can represent a spatial relationship (e.g., “adjoins,” “supports,” “contains,” etc.). In these embodiments, each node within the graph database represents a digital twin of an entity (e.g., a shipping entity) and may include data structures defining the digital twin. In these embodiments, each edge within the graph database may represent a relationship between two entities represented by connected nodes. In some implementations, an edge may represent a spatial relationship (e.g., “adjoins,” “supports,” “interlocks,” “supports,” “contains,” etc.). In embodiments, various types of data may be stored in nodes or edges. In embodiments, a node may store attribute data, state data, and/or metadata associated with a facility, system, subsystem, and/or component. The types of attribute data and state data will be different based on the entity represented by the node. For example, a node representing a delivery robot may include attribute data representing the material of the robot, the dimensions of the robot (or its components), the mass of the robot, etc. In this example, the robot's state data may include the robot's current pose, the robot's location (e.g., within a smart container or on a container ship), etc. In embodiments, an edge may store relationship data and metadata data regarding the relationship between two nodes. Examples of relationship data may include the nature of the relationship, whether the relationship is persistent (e.g., a fixed component will have a permanent relationship with the structure to which it is attached or supported), etc. In embodiments, an edge may include metadata regarding a relationship between two entities. For example, a sensor may perform measurements related to the state of a smart container, whereby one relationship between a sensor and a smart container may include "measured" and define the type of measurement measured by the sensor. You can. In this example, metadata stored at the edge may include a list of N measurements taken and a timestamp of each individual measurement. In this way, temporal data about the nature of the relationship between two entities can be maintained, thereby facilitating cause-and-effect analysis used in, for example, predictive systems, analytics engines, and machine learning. Allow the engine, and/or visualization engine, to utilize such temporal relationship data, such as by aligning disparate data sets with a series of viewpoints.

In some embodiments, a graph database may be implemented in a hierarchical manner such that the graph database is associated with a set of facilities, systems, and components. For example, a digital twin of a delivery environment may include nodes representing the delivery environment. The graph database may further include nodes representing various systems within the delivery environment, such as nodes representing a smart container fleet, smart container charging area, storage area, etc., all of which are nodes representing the delivery environment. can be connected to In this example, each of the systems may be further connected to various subsystems and/or components of the system. For example, a smart container system may include a subsystem node representing the cooling system of the smart container, a second subsystem node representing the heating system of the smart container, a third subsystem node representing the fan system of the smart container, and the thermostat of the smart container. It may be connected to one or more nodes representing a thermostat (or multiple thermostats). With further reference to this example, subsystem nodes and/or component nodes may be connected to lower level nodes that may include subsystem nodes and/or component nodes. For example, a subsystem node representing a cooling subsystem may be connected to a component node representing an air conditioning unit. Similarly, a component node representing a thermostat device may be connected to one or more component nodes representing various sensors (eg, temperature sensors, humidity sensors, etc.).

In embodiments where a graph database is implemented, the graph database may be associated with a single environment or may represent a larger enterprise. In the latter scenario, the company may have multiple shipping distribution facilities. In these embodiments, an enterprise node representing an enterprise may connect to each facility's environment nodes. In this way, the digital twin module 13420 can maintain a digital twin for a company's multiple delivery facilities.

In embodiments, digital twin module 13420 may use a graph database to create a digital twin that can be rendered and displayed and/or represented as a data representation. In embodiments, digital twin module 13420 may receive a request to render a digital twin, whereby the request includes one or more parameters indicating the view to be depicted. For example, one or more parameters may determine the type of delivery environment and rendering to be depicted (e.g., a “real world view” depicting the environment as seen by humans, an “infrared view” depicting objects as a function of their respective temperatures). , an “airflow view” depicting airflow in the digital twin, etc.). In response, the digital twin module 13420 examines the graph database and determines the nodes in the graph database that are related (either directly or through low-level nodes) to an environment node in the environment and the edges that define the relationships between the related nodes. Based on these, the composition of the environment to be depicted can be determined. When determining the configuration, digital twin module 13420 can identify the surfaces to be depicted and render these surfaces. Digital twin module 13420 can then render the requested digital twin by connecting the surfaces according to their configuration. The rendered digital twin can then be output to a viewing device (e.g., VR headset, AR headset, monitor, etc.). In some scenarios, digital twin module 13420 may receive real-time sensor data from a sensor system in the environment and update the visual digital twin based on the sensor data. For example, digital twin module 13420 may receive sensor data related to smart container cargo (e.g., vibration data from a vibration sensor). Based on the sensor data, digital twin module 13420 can update the visual digital twin to represent approximate vibration characteristics of the cargo within the smart container's digital twin.

In scenarios where digital twin module 13420 is providing data representations of digital twins (e.g., for dynamic modeling, simulations, machine learning), digital twin module 13420 can delve into the graph database. and determine the configuration of the environment to be depicted based on nodes in the graph database that are related (either directly or through low-level nodes) to an environment node in the environment and edges that define relationships between related nodes. In some scenarios, digital twin module 13420 may receive real-time sensor data from sensor systems in the delivery entity and/or environment and apply one or more dynamic models to the digital twin based on the sensor data. In other scenarios, the digital twin's data representation can be used to perform simulations, as discussed in more detail throughout this specification.

In some embodiments, digital twin module 13420 may execute digital ghosts that execute against digital twins of shipping entities (e.g., smart container fleets or individual smart containers) and/or shipping environments. In these embodiments, the digital ghost may monitor one or more sensors of the sensor system of the delivery entity and/or the environment to detect anomalies that may indicate malicious viruses, compromised sensors, or other security issues.

As discussed, the digital twin module 13420 may include the digital twin management system 13306, the digital twin I/O system 13308, the digital twin simulation system 13310, the digital twin dynamic model system 13312, and/or the digital twin management system 13306. May include twin control module 13314.

In embodiments, digital twin management system 13306 creates new digital twins, maintains/updates existing digital twins, and/or renders digital twins. Digital twin management system 13306 may receive user input, uploaded data, and/or sensor data to create and maintain existing digital twins. When creating a new digital twin, digital twin management system 13306 may store the digital twin in digital twin data store 13316. Creating, updating, and rendering digital twins are discussed in greater detail throughout this disclosure.

In embodiments, digital twin I/O system 13308 receives input from various sources and outputs data to various recipients. In embodiments, the digital twin I/O system receives sensor data from one or more sensor systems. In these embodiments, each sensor system may include one or more sensors outputting respective sensor data. Each sensor may be assigned an IP address or other suitable identifier. Each sensor may output sensor packets including the sensor identifier and sensor data. In some embodiments, sensor packets may further include a timestamp indicating when the sensor data was collected. In some embodiments, digital twin I/O system 13308 may interface with a sensor system through a real-time sensor API. In these embodiments, one or more devices (e.g., sensors, aggregators, edge devices) of the sensor system transmit sensor packets containing sensor data to the digital twin I/O system 13308 via an API. can do. The digital twin I/O system can determine the sensor system that transmitted the sensor packet and its cargo, and manages the sensor data and any other related data (e.g., timestamps, environmental identifiers/sensor system identifiers, etc.) as a digital twin. It can be provided to the system 13306.

In embodiments, digital twin I/O system 13308 may receive imported data from one or more sources. For example, the digital twin module 13420 may provide a portal for users to create and manage digital twins. In these embodiments, a user may upload one or more files (e.g., image files, LIDAR scans, blueprints, etc.) associated with the new digital twin being created. In response, digital twin I/O system 13308 may provide the imported data to digital twin management system 13306. Digital twin I/O system 13308 may receive other suitable types of data without departing from the scope of this disclosure.

In some embodiments, digital twin simulation system 13310 is configured to run simulations using a digital twin. For example, digital twin simulation system 13310 may iteratively adjust one or more parameters of the digital twin and/or one or more embedded digital twins. In embodiments, digital twin simulation system 13310 may, for each set of parameters, run a simulation based on the set of parameters and collect simulation result data resulting from the simulation. In other words, the digital twin simulation system 13310 may collect data regarding the properties of the digital twin and the digital twin within or including the digital twin used during the simulation, as well as any results resulting from the simulation. For example, when running a simulation for a digital twin of a potential new smart container design, digital twin simulation system 13310 may change dimensions, materials, capabilities, and/or other relevant parameters, Simulations can be run that output results arising from different combinations. As another example, the digital twin simulation system 13310 can simulate the vibration of cargo within a smart container. In this example, the smart container's digital twin may include a set of operating parameters of the smart container. In some embodiments, operating parameters may be changed to evaluate their effect on cargo damage. Digital twin simulation system 13310 is discussed in more detail throughout this disclosure.

In an embodiment, digital twin dynamic model system 13312 is configured to model one or more behaviors on a digital twin of the environment. In embodiments, digital twin dynamic model system 13312 may receive a request to model a particular type of behavior regarding a delivery entity, environment, or process, and create a dynamic model, a digital twin of the environment or process, and an environment or process. Its behavior can be modeled using sensor data collected from one or more sensors that are monitoring. For example, the operator of a smart container fleet may want to model the performance of the fleet to determine whether the fleet can withstand increased demand for cargo transportation services. In this example, digital twin dynamic model system 13312 may execute a dynamic model configured to determine whether an increase in demand will result in negative consequences (e.g., failures, downtime, etc.). Digital twin dynamic model system 13312 is discussed in more detail throughout this disclosure.

In embodiments, intelligence service 13004 performs machine learning and artificial intelligence-related tasks on behalf of the digital twin system. In an embodiment, intelligence service 13004 uses output of simulations run by digital twin simulation system 13310 to train a machine learning model. In some of these embodiments, the results of simulations may be used to supplement training data collected from real-world environments and/or processes. In embodiments, intelligence service 13004 utilizes machine learning models to perform predictions and classifications and provide decision support regarding real-world environments and/or processes represented by each digital twin.

For example, a machine-learned predictive model could be used to predict the cause of irregular vibration patterns in the bearings of a smart container's engine. In this example, intelligence service 13004 may receive vibration sensor data from one or more vibration sensors placed on or near the engine, may receive maintenance data from a smart container, and may receive vibration sensor data and maintenance data from a smart container. A feature vector can be created based on data. Intelligence service 13004 inputs the feature vectors into a machine learning model specifically trained on the engine (e.g., using simulated data in combination with real data of the causes of irregular vibration patterns) to predict the causes of irregular vibration patterns. can do. In this example, the cause of the irregular vibration pattern could be loose bearings, lack of bearing lubricant, misaligned bearings, worn bearings, the bearing's phase being out of alignment with the engine's phase, loose housing, loose bolts, etc.

In embodiments, digital twin control module 13314 controls one or more aspects of smart containers, smart container fleets, and/or other shipping entities and environments. In embodiments, digital twin control module 13314 may utilize digital twin simulation system 13310, digital twin dynamic model system 13312, and/or intelligence services 13004 to determine one or more control commands. In embodiments, digital twin control module 13314 may implement rule-based and/or machine learning approaches to determine control commands. In response to determining control commands, digital twin control module 13314 may output control commands to smart containers, smart container fleets, and/or other shipping entities and environments via digital twin I/O system 13308. .

In embodiments, the digital twin management system 13306 may include, but is not limited to, a digital twin configuration module 13318, a digital twin update module 13320, and a digital twin visualization module 13322.

In embodiments, the digital twin configuration module 13318 may configure input from users, imported data (e.g., blueprints, specifications, etc.), image scans of the environment, 3D data from LIDAR devices and/or SLAM sensors. , and other appropriate data sources can be used to construct a new set of digital twins of the set of environments. For example, a user (e.g., a user affiliated with an organization/customer account) may input input to create a new shipping digital twin via client application 13324 (e.g., via smart container system 13000). can be provided. In doing so, users may upload 2D or 3D image scans and/or blueprints of the delivery entity and/or environment. Users can also upload 3D data captured by, for example, a camera, LIDAR device, IR scanner, set of SLAM sensors, radar device, EMF scanner, etc. In response to the provided data, digital twin construction module 13318 may generate a 3D representation of the delivery entity or environment, which may include any objects captured in the image data captured/detected in the 3D data. In embodiments, intelligence service 13004 may analyze input data (e.g., blueprints, image scans, 3D data) to classify rooms, paths, equipment, etc. to assist in the creation of 3D representations. In some embodiments, digital twin configuration module 13318 may map the digital twin to a 3D coordinate space (e.g., Cartesian space with x, y, and z axes).

In some embodiments, digital twin configuration module 13318 may output a 3D representation of the delivery entity and/or environment to a graphical user interface (GUI). In some of these embodiments, a user can identify specific areas and/or objects and provide input regarding the identified areas and/or objects. For example, a user can label specific rooms, equipment, machines, devices, sensors, etc. Additionally or alternatively, the user may provide data regarding the identified objects and/or areas. For example, in identifying a smart container, a user may provide the smart container's manufacturer/model number. In some embodiments, digital twin configuration module 13318 may obtain information from the manufacturer of the device, equipment, or machine. This information may include one or more properties and/or behavior of the device, equipment, or machine. In some embodiments, a user may identify the locations of sensors throughout the delivery entity and/or environment via a GUI. For each sensor, the user can provide the sensor's type and related data (e.g., manufacturer, model, IP address, etc.). Digital twin configuration module 13318 may record the location of the delivery digital twin (e.g., x, y, z coordinates of sensors). In embodiments, digital twin module 13420 may utilize one or more systems to automate a population of digital twins. For example, digital twin module 13420 may employ a machine vision-based classifier to classify the make and model of a device, equipment, or sensor. Additionally or alternatively, digital twin module 13420 may repeatedly ping different types of known sensors to identify the presence of specific types of sensors in the environment. Whenever a sensor responds to a ping, digital twin module 13420 can extrapolate the sensor's make and model.

In some embodiments, manufacturers may provide or make available digital twins of their products (e.g., sensors, devices, machines, equipment, raw materials, etc.). In these embodiments, digital twin configuration module 13318 can import digital twins of one or more products identified in the environment and embed these digital twins in the digital twin of the environment. In embodiments, embedding a digital twin within another digital twin may include creating a relationship between the embedded digital twin and the other digital twin. In these embodiments, the manufacturer of the digital twin can define the behaviors and/or properties of individual products. For example, the digital twin of a 3D printer within a smart container can define how the 3D printer operates, the input/output of the 3D printer, etc. In this way, a digital twin of a 3D printer can reflect the behavior of the 3D printer given a set of inputs.

In embodiments, a user may define one or more delivery processes. In these embodiments, a user can define the steps of the process, the machines/devices that perform each step of the process, the inputs to the process, and the outputs of the process.

In embodiments, digital twin construction module 13318 may create a graph database defining relationships between sets of digital twins. In these embodiments, digital twin configuration module 13318 may configure a delivery entity and/or environment, systems and subsystems of the delivery entity and/or environment, devices within the delivery entity and/or environment, and Nodes can be created for sensors within the environment, workers working in the delivery environment, delivery processes performed involving the delivery entity and/or environment, etc. In embodiments, digital twin configuration module 13318 may write a graph database representing a set of digital twins to digital twin data store 13316.

In embodiments, digital twin configuration module 13318 may include, for each node, any data about the entity within the node that represents the entity. For example, when defining a node representing a container ship, digital twin configuration module 13318 may include dimensions, boundaries, layout, routes, and other relevant spatial data at the node. Additionally, the digital twin configuration module 13318 may define a coordinate space for the container ship. If the digital twin can be rendered, the digital twin configuration module 13318 can include references within the node to any geometry, mesh, spline, surface, etc. that can be used to render the environment. In representing a system, subsystem, device, or sensor, digital twin configuration module 13318 can create a node for each entity and include any associated data. For example, digital twin configuration module 13318 can create a node representing a delivery robot. In this example, digital twin configuration module 13318 may include dimensions, behaviors, properties, location, and/or any other suitable data about the robot at the node representing the robot. Digital twin configuration module 13318 may create relationships between entities by connecting nodes of related entities to edges. In doing so, the created relationships between entities can define the types of relationships characterized by edges. When representing a process, digital twin configuration module 13318 can create a node for the entire process or create nodes for each step of the process. In some of these embodiments, digital twin configuration module 13318 may relate process nodes to nodes representing machinery/devices that perform steps in the process. In embodiments where an edge connects process step nodes to the machinery/device that performs the process step, one of the edges or nodes may be connected to the input to the step, the output of the step, the amount of time the step takes, and the input to the step to produce the outputs. It may contain information indicating the nature of the processing of inputs, the set of states or modes that the process may experience, etc.

In an embodiment, digital twin update module 13320 updates a set of digital twins based on the current state of one or more delivery entities and/or the environment. In some embodiments, digital twin update module 13320 receives sensor data from a sensor system in the delivery entity and/or environment, and updates the digital twin of the delivery entity or environment and/or any affected systems, subsystems, devices, Update the status of the digital twin, including workers and processes. As discussed, digital twin I/O system 13308 may receive sensor data in one or more sensor packets. Digital twin I/O system 13308 can provide sensor data to digital twin update module 13320 and identify the entity or environment from which the sensor packet was received and the sensor that provided the sensor packet. In response to sensor data, digital twin update module 13320 may update the status of one or more digital twins based on the sensor data. In some of these embodiments, digital twin update module 13320 may update records (e.g., nodes in a graph database) corresponding to sensors that provided sensor data to reflect current sensor data. In some scenarios, digital twin update module 13320 may identify specific areas within the environment or entities being monitored by sensors and update records (e.g., nodes within a graph database) to reflect current sensor data. can do. For example, digital twin update module 13320 may receive sensor data reflecting different vibration characteristics of the smart container and/or components. In this example, digital twin update module 13320 may update records representing the vibration sensors that provided the vibration sensor data and/or records representing the smart container and/or smart container components to reflect the vibration sensor data. . In another example, in some scenarios, workers in a shipping environment (e.g., container port, container storage facility, etc.) wear wearable devices (e.g., smart watches, smart helmets, smart shoes, etc.) It may be required to do so. In these embodiments, wearable devices may collect sensor data (e.g., location, movement, heart rate, breathing rate, body temperature, etc.) about the worker and/or the environment surrounding the worker, and may display the collected sensor data (e.g. It may communicate to the digital twin module 13420 directly (for example, via the real-time sensor API 13326) or via an aggregation device in the sensor system. In response to receiving sensor data from the worker's wearable device, digital twin update module 13320 may update the worker's digital twin to reflect, for example, the worker's location, the worker's trajectory, the worker's health status, etc. You can. In some of these embodiments, digital twin update module 13320 may update edges connecting nodes representing the worker and/or nodes representing the environment with collected sensor data to reflect the worker's current state.

In some embodiments, digital twin update module 13320 may provide sensor data from one or more sensors to digital twin dynamic model system 13312, which may use the delivery environment and/or one or more The behavior of delivery entities can be modeled.

In an embodiment, digital twin visualization module 13322 receives a request to view a visual digital twin or portion thereof. In embodiments, the request may indicate a digital twin (e.g., smart container identifier) to be viewed. In response, digital twin visualization module 13322 may determine the requested digital twin and any other digital twins involved by the request. For example, when requesting to view a digital twin of a smart container, digital twin visualization module 13322 may further identify a digital twin of any shipping entity within the smart container. In embodiments, digital twin visualization module 13322 may identify spatial relationships between shipping entities and smart containers, for example, based on relationships defined in a graph database. In these embodiments, the digital twin visualization module 13322 may display the relative positions of embedded digital twins within the containing digital twin, the relative positions of adjacent digital twins, and/or the transient state of the relationship (e.g., It can be determined whether the object is fixed at one point or the object is moving. Digital twin visualization module 13322 may render the requested digital twins and any other nested digital twins based on the identified relationships. In some embodiments, digital twin visualization module 13322 can determine, for each digital twin, the surfaces of the digital twin. In some embodiments, surfaces of the digital may be defined or referenced in a record corresponding to the digital twin, which may be provided by the user, determined from imported images, or defined by the manufacturer of the shipping entity. In scenarios where an object may assume different poses or shapes (e.g., a delivery robot), digital twin visualization module 13322 may determine the pose or shape of the object for the digital twin. The digital twin visualization module 13322 can embed the digital twin in the requested digital twin and output the requested digital twin to the client application.

In some of these embodiments, a request to view a digital twin may further indicate the type of viewing. As discussed, in some embodiments, digital twins may be shown in multiple different view types. For example, a delivery entity or environment can be divided into a "real-world" view that depicts the environment or device as they typically appear, a "heat" view that depicts the environment or entity in a way that represents the temperature of the environment or entity, a “vibration” view that depicts shipping entities in a way that represents the vibrational properties of the entities, a “velocity” view that depicts shipping entities in a way that represents the velocity of the entities, and a “velocity” view that depicts shipping entities in a way that represents the oscillation characteristics of the entities, a “velocity” view that depicts shipping entities in a way that represents the oscillation characteristics of the entities, and a “velocity” view that depicts shipping entities in a way that represents the oscillation characteristics of the entities. a “filtered” view that displays only objects requiring attention (resulting from recognition of fault conditions, warnings, updated reports, or other factors), an augmented view that overlays data on the digital twin, and/or Can be displayed in any other suitable view types.

In embodiments, digital twins may be depicted in multiple different role-based view types. For example, a smart container fleet may be viewed in a "manager" view, which depicts smart containers in a manner suitable for a smart container fleet manager, and a container terminal environment may be viewed in a "manager" view, which depicts container terminals in a manner suitable for a container terminal operator. Smart containers can be viewed in the “Operator” view, which depicts the facility in a manner appropriate for the regulatory administrator, and Shippers can be viewed in the “shipper” view, which depicts the smart container in a manner appropriate for the shipper. You can watch the container digital twin (13504). In response to a request indicating a view type, digital twin visualization module 13322 may retrieve data for each digital twin corresponding to the view type. For example, if a user requested a hit view of a smart container fleet, the digital twin visualization module 13322 may display (smart containers, shipping environments, temperature measurements taken from different smart container components, and/or digital twin dynamics). Temperature data for that set of smart containers (which may include extrapolated temperature measurements by model system 13312 and/or simulated temperature data from digital twin simulation system 13310) as well as any other shipping entity. You can search available temperature data for: In this example, digital twin visualization module 13322 displays colors corresponding to each smart container that indicate temperature fault level status (e.g., red for alert, orange for threshold, yellow for suboptimal). , and green for normal operation) can be determined. The digital twin visualization module 13322 can then render digital twins of smart containers based on the determined colors. Note that in some embodiments, digital twin module 13420 may include an analysis system (not shown) that determines how digital twin visualization module 13322 presents information to a human user. For example, analytics systems can track outcomes related to human interactions with real-world environments or objects in response to information presented in a visual digital twin. In some embodiments, the analysis system can visualize information based on the resulting data (e.g., what colors to use to indicate an alarm condition, what types of movements or animations to draw attention to an alarm condition, etc.). Alternatively, cognitive models can be applied to determine the most effective way to display audio information (what sounds to use to indicate alarm conditions). In some embodiments, the analysis system may apply cognitive models to determine the most appropriate way to display visualized information based on the user's role. In embodiments, the visualization may include graphical information, graphical information depicting physical characteristics, graphical information depicting financial characteristics, graphical information depicting performance characteristics, recommendations from intelligence service 13004, intelligence service 13004 Predictions from, probability of failure data, maintenance history data, time-to-failure data, downtime cost data, probability of downtime data, repair cost data, replacement cost (e.g., probability of replacing a smart container or a smart container component) It may include the display of information related to the visualized digital twins, including the like.

In an embodiment, smart container fleet manager digital twin 13328 is a digital twin configured for managers and/or operators of smart container fleets. In embodiments, smart container fleet manager digital twin 13328 operates in conjunction with system 13000 to perform analytics, machine learning, and/or other AI and input (e.g., maritime data, traffic data, weather data). , sensor data, regulatory data, etc.) may provide simulation, optimization, classification, configuration and/or control, prediction, statistical summarization, and decision support. In an embodiment, smart container fleet manager digital twin 13328 is responsible for confirming cargo storage and/or transportation service orders, selecting smart container modes of transportation/route, and maintaining service transactions with third-party service providers. engages in, inspects individual smart containers, monitors smart container fleets, creates smart contracts, monitors regulatory compliance, performs risk management, and other fleet manager-related activities. Functions that are not limited to this may be provided.

In embodiments, the types of data that may populate smart container fleet manager digital twin 13328 include: financial data, weather data, macroeconomic data, microeconomic data, forecast data, demand planning data, AI and/or machine learning. Analytical results of modeling (e.g. financial forecasting), forecast data, asset data, recommendation data, strategic competitive data (e.g. news and events about industry trends and competitors), transportation data, maritime data , trucking data, freight data, airline data, rail data, traffic data, social media data, survey data, and many others. In embodiments, digital twin module 13420 may obtain financial data from, for example, publicly disclosed financial statements, third-party reports, tax returns, public news sources, etc. In embodiments, macroeconomic data may be derived by analyzing various financial and operational data collected by system 13000. In embodiments, business performance metrics may be analytically derived by intelligence service 13004, based at least in part on real-time behavioral data and/or provided from other users and/or their respective trader digital twins.

In embodiments, smart container fleet manager digital twin 13328 provides high-level views of different states of the fleet, real-time representations of the fleet, historical representations of the fleet, and projected representations of the fleet (e.g., future states). s), real-time representations of individual smart containers, historical representations of individual smart containers, projected representations (e.g., future states) of individual smart containers, real-time representations of shippers, historical representations of shippers, shippers projected representations of shipping lines, real-time representations of shipping lines, historical representations of shipping lines, projected representations of shipping lines, news and/or television data, economic sentiment data, social media It can include data, charts, countdowns to closing information, lease terms, smart contract terms, contract terms, and many other things. In embodiments, smart container fleet manager digital twin 13328 may allow users to access and/or interact with other shipping digital twins. In embodiments, smart container fleet manager digital twin 13328 allows a user to access and/or interact with a fleet of smart container digital twins 13504 and/or individual smart container digital twins 13504. can do. In embodiments, smart container fleet manager digital twin 13328 may allow users to interact with other smart container fleet manager digital twins 13328 and/or shipper digital twins 13502. The smart container fleet manager digital twin 13328 may initially depict various states at a lower level of granularity. In embodiments, a user viewing smart container fleet manager digital twin 13328 can drill down into selected states and choose to view the selected states at a higher level of granularity. For example, the Smart Container Fleet Manager Digital Twin 13328 initially monitors a subset of the various states of a Smart Container Fleet, including the price status (e.g., a visual indicator indicating the price for a Smart Container) at a lower granularity. It can be shown in levels. In response to your selection, the smart container fleet manager digital twin 13328 provides real-time, historical, aggregated, comparative, and/or predicted pricing data (e.g., real-time, historical, simulated, and/or predicted revenue, liabilities, etc.). may provide data, analysis, summaries and/or reporting, including but not limited to; In an embodiment, the smart container fleet manager digital twin 13328 initially provides a user (e.g., a fleet manager) with views of various different aspects of the fleet (e.g., representing different “health” levels of the fleet). different indicators), but allow the user to choose which aspect requires more attention from the user. In response to this selection, smart container fleet manager digital twin 13328 may request a more granular view of the selected state(s) from system 13000, which may return the requested states at a more granular level. .

In embodiments, digital twin simulation system 13310 may receive a request to perform a simulation requested by smart container fleet manager digital twin 13328, where the request specifies one or more parameters to be changed in one or more digital twins. indicates. In response, digital twin simulation system 13310 may return simulation results to smart container fleet manager digital twin 13328, which ultimately outputs the results to the user via a client device display. In this way, the user can be provided with various results corresponding to different parameter configurations. For example, a user may request a set of simulations to be run to test different fleet configurations to see how the different configurations affect the overall impact on profits and losses. The digital twin simulation system 13310 can simulate varying different configurations and output financial forecasts for each individual configuration. In some embodiments, a user can select a parameter set based on various results and repeat simulations based on at least changed previous results. In some embodiments, an intelligent agent may be trained to recommend and/or select parameter sets based on respective results associated with each individual parameter set.

In embodiments, smart container fleet manager digital twin 13328 stores, aggregates, merges, and analyzes material related to pricing, scheduling, financial reporting, performance, maintenance, regulatory data, or other data related to smart container shipping services. It can be configured to prepare, report, and distribute. As described herein, smart container fleet manager digital twin 13328 can link to, interact with, and associate with external data sources, and can access external data sources, including data internal to system 13000. You can upload, download, aggregate sources, and analyze such data. Data analytics, machine learning, AI processing, and other analytics may be coordinated between the smart container fleet manager digital twin 13328 and intelligent services 13004. This collaboration and interaction can be used in modeling, machine learning, and AI processing to identify optimal fleet configuration, optimal scheduling of cargo storage and/or delivery service orders, or some other delivery-related metric or aspect. This may include seeding delivery-related data elements and domains into the twin data store 13316, as well as assisting in the determination of fleet composition or identification of optimal data measurement parameters based on scheduling execution success. there is. In embodiments, digital twin module 13420 abstracts different views (or states) within the digital twin to appropriate granularity. For example, digital twin module 13420 may have access to real-time sensor data streams as well as access to all sensor data collected on behalf of 13000. In this example, if sensor readings from a particular smart container indicate a potentially critical situation (e.g., malfunction, hazardous condition, damaged cargo, potentially illegal cargo, etc.), the analysis indicates a potentially critical situation. can become very important to fleet managers. Accordingly, digital twin module 13420 may include status indicators of smart containers in fleet manager digital twin 13328 when building an appropriate view for the fleet manager. In this way, fleet managers can drill down into the health indicators of smart containers to view potentially critical situations at a more granular level (e.g., analysis of sensor data used to identify situations and smart container machines).

In an embodiment, smart container fleet manager digital twin 13328 may be configured to report on the performance of smart containers within a fleet. As described herein, reporting may include timing performance metrics, financial performance metrics, physical performance metrics, cargo damage metrics, data regarding resource usage, or some other type of reporting data. In embodiments, an intelligent agent trained by a user may be trained to surface the reports that are most important to the user. For example, if a user (e.g., a fleet manager) continuously views and carefully observes timing performance, but routinely skips reporting related to financial performance, the execution agent may suppress financial performance data while also Reports can be automatically surfaced to users.

In embodiments, smart container fleet manager digital twin 13328 monitors, stores, aggregates, merges, analyzes, and prepares materials associated with other shipping entities (e.g., shippers, shipping lines, container terminals, or named entities of interest). , reporting and distribution. In embodiments, such data may be used to retrieve and collect shipping entity information from sources including, but not limited to, regulatory information, information regarding shipping, press releases, SEC or other financial reports, or some other publicly available data. Data may be collected by system 13000 through aggregation, webscraping, or other techniques. For example, a user wishing to monitor a specific shipping entity could request that the smart container fleet manager digital twin 13328 provide material regarding that specific shipping entity. In response, system 13000 may identify a set of data sources (e.g., internal data sources, authorized third party data, etc.) that are publicly available or accessed by the fleet manager.

In embodiments, a client application 13324, such as 13000 running a smart container fleet manager digital twin 13328, may be configured as an intelligent agent trained on the fleet manager's actions (which may indicate behavior and/or preferences). there is. In embodiments, the intelligent agent may record actions and associated characteristics (e.g., context associated with the user's actions) into the intelligent agent system. For example, the Intelligent Agent will be able to monitor each time a user approves a freight storage and/or delivery service order (which is an action), as well as characteristics surrounding the approval (e.g., type of action, type of order, price of the order, Shipper, quantity of smart container, route information, etc.) can be recorded. The intelligent agent may report actions and characteristics to the intelligent agent system that can train the intelligent agent on how the intelligent agent may undertake or recommend approval tasks and other tasks in the future. Once trained, the intelligent agent can automatically perform actions and/or recommend actions to the user. Additionally, in embodiments, the intelligent agent may record results related to performed/recommended actions, thereby creating a feedback loop with the intelligent agent system.

In embodiments, smart container fleet manager digital twin 13328 may provide an interface for a fleet manager to perform one or more fleet manager-related workflows. For example, the Smart Container Fleet Manager Digital Twin (13328) allows managers to configure cargo transportation order approval workflows, smart container maintenance workflows, logistics workflows, smart contract workflows, shipping and/or delivery workflows, and regulatory workflows. An interface may be provided to perform, supervise, or monitor, etc.

In another example, a user may request a filtered view of the digital twin of a process, whereby the digital twin of the process shows only the delivery entities involved in the process. In this example, digital twin visualization module 13322 retrieves the digital twin of the process as well as any related digital twins (e.g., the digital twin of the environment and digital twins of any shipping entities that affect the process). can do. Digital twin visualization module 13322 may then render each of the digital twins (e.g., the environment and associated delivery entities) and then perform processes on the rendered digital twins. Note that because a process may be performed over a period of time and may include moving items and/or parts, digital twin visualization module 13322 may generate a series of sequential frames that demonstrate the process. . In this scenario, the movements of the delivery entities involved by the process may be determined according to the behaviors defined in the respective digital twins of the machines and/or devices.

As discussed, digital twin visualization module 13322 may output the requested digital twin to client application 13324. In some embodiments, client application 13324 is a virtual reality application whereby the requested digital twin is displayed on a virtual reality headset. In some embodiments, client application 13324 is an augmented reality application, whereby the requested digital twin is displayed on an AR-enabled device. In these embodiments, the requested digital twin may be filtered such that visual elements and/or text are overlaid on the display of the AR-enabled device.

Note that although a graph database is discussed, digital twin module 13420 may utilize other suitable data structures to store information about the set of digital twins. In these embodiments, the data structures and any associated storage system may be implemented to provide some degree of feedback loops and/or recursion when the data structures represent repetition of flows.

In embodiments, the digital twin I/O system 13308 may be configured to connect a shipping entity and/or environment, digital twin module 13420, to provide bi-directional transfer of data between coupled components in accordance with some embodiments of the present disclosure. ), and/or interfaces with its components.

In embodiments, the transmitted data may include signals between connected components, which may include software components, hardware components, physical devices, virtualized devices, simulated devices, combinations thereof, etc. For example, request signals, command signals, response signals, etc.). Signals can be material properties (e.g. physical quantities such as temperature, pressure, humidity, density, viscosity, etc.), measured values (e.g. simultaneous or stored values obtained by a device or system), device properties (e.g. For example, properties of the device ID or design specifications of the device, materials, measurement capabilities, dimensions, absolute position, relative position, combinations thereof, etc.), set points (e.g. targets for material properties, device properties, system properties, combinations thereof, etc.), and/or threshold points (e.g., minimum or maximum values for material properties, device properties, system properties, etc. Thresholds such as ) can be defined. Signals may be received from systems or devices that acquire (e.g., directly measure or generate) or directly obtain (e.g., receive, compute, lookup, filter, etc.) data, and may be received at a predetermined time or digitally. May be communicated to or from digital twin I/O system 13308 in response to a request (e.g., polling) from twin I/O system 13308. Communication may occur through direct or indirect connections (eg, through intermediate modules within the circuit and/or intermediate devices between connected components). The values may correspond to real-world elements or virtual elements (e.g., input or output to a digital twin providing data and/or a simulated element).

In embodiments, real-world elements may be delivery entities or elements within the environment. Real-world elements may include, for example, non-network elements, devices (smart or non-smart), sensors, and humans. Real-world elements may be shipping entities or process or non-process equipment within the environment. For example, process equipment may include motors, cranes, reach stackers, forklifts, pumps, fans, etc., while non-process equipment may include personal protective equipment, safety equipment, emergency stations or devices (e.g. , safety showers, eyewashes, fire extinguishers, sprinkler systems, etc.), container terminals or other facility features (e.g., walls, floor layout, etc.), obstacles (e.g., people or other goods within the entity or environment), etc. It can be included.

Virtual elements may be digital representations of, or correspond to, simultaneously existing real-world elements. Additionally or alternatively, virtual elements may be digital representations of or corresponding real-world elements that may be available for later addition and implementation into an entity or environment. Virtual elements may include, for example, simulated elements and/or digital twins. In embodiments, simulated elements may be digital representations of real-world elements that do not exist within the delivery entity or environment. Simulated elements can mimic desired physical properties that can later be incorporated within an entity or environment as real-world elements (e.g., a “black box” that mimics the dimensions of real-world elements). Simulated elements may include digital twins of existing objects (e.g., a single simulated element may include one or more digital twins for existing sensors). Information related to simulated elements is mapped using mathematical models or algorithms, for example from libraries defining the information and behavior of the simulated elements (e.g. physical libraries, chemical libraries, etc.). It can be obtained by evaluating the behavior of real-world elements.

In embodiments, a digital twin may be a digital representation of one or more real-world elements. Digital twins are configured to mimic, copy, and/or model the behaviors and responses of real-world elements in response to inputs, outputs, and/or conditions of the surrounding environment. Data related to the physical properties and responses of real-world elements can be obtained, for example, from user input, sensor input, and/or physical modeling (e.g., thermodynamic models, electrodynamic models, mechanistic models, etc.). It can be obtained through Information about the digital twin may correspond to and be obtained from one or more real-world elements that correspond to the digital twin. For example, in some embodiments, the digital twin may correspond to one real-world element, a digital vibration sensor fixed on a smart container cargo, and vibration data for the digital twin can be transmitted to a digital vibration sensor fixed on the cargo. It can be obtained by polling or fetching the vibration data measured by. In a further example, the digital twin may correspond to a plurality of real-world elements that are fixed digital vibration sensors on a smart container component, and the vibration data for the digital twin includes vibration data measured by each of the fixed digital vibration sensors on the plurality of real-world elements. Can be obtained by polling or fetching. Additionally or alternatively, the vibration data of the first digital twin may be obtained by fetching the vibration data of a second digital twin embedded in the first digital twin, and the vibration data for the first digital twin may be acquired by fetching the vibration data of the second digital twin embedded in the first digital twin. It may include or be derived from vibration data. For example, the first digital twin may be a digital twin of a smart container, and the second digital twin may be a digital twin corresponding to the cargo within the smart container, such that data including vibration data for the second digital twin is derived from or thereto. Based on this, vibration data for the first digital twin is obtained or calculated.

In embodiments, digital twin module 13420 monitors properties of real-world elements using sensors, which can be represented by a digital twin and/or outputs of models for one or more simulated elements. In embodiments, digital twin module 13420 may use data acquired from other sources (e.g., sensors physically proximate to or influencing the affected real-world elements) for extended intervals (e.g. Maintain effective monitoring of processes by performing simulations (e.g., via a digital twin simulation system 13310) and by minimizing data transmission to sensors corresponding to affected real-world elements and/or extending polling intervals. This can minimize network congestion. Additionally or alternatively, error checking may be performed by comparing collected sensor data to data obtained from digital twin simulation system 13310. For example, consistent deviations or fluctuations between sensor data obtained from real-world elements and simulated elements may indicate a malfunction or other fault condition of the respective sensor.

In embodiments, digital twin module 13420 may optimize features of smart container fleets, smart containers, and other shipping entities and/or environments through the use of one or more simulated elements. For example, digital twin module 13420 may evaluate the effectiveness of simulated elements within a digital twin of a smart container to quickly and efficiently determine the costs and/or benefits flowing from the inclusion, exclusion, or replacement of real-world elements within the smart container. there is. Costs and benefits include, for example, manufacturing costs, maintenance costs, efficiency (e.g. process optimization to reduce waste or increase throughput), climate considerations (e.g. carbon footprint), and longevity. , minimization of component defects, component downtime, etc.

In embodiments, the digital twin I/O system 13308 may be configured to control one or more devices (e.g., server devices, user devices, and/or distributed devices) to effect the described functions. It may include one or more software modules executed by one or more controllers. Digital twin I/O system 13308 may include, for example, input modules, output modules, and adapter modules.

In embodiments, input modules may acquire or import data from data sources that communicate with digital twin I/O system 13308, such as sensor systems and digital twin simulation system 13310. Data can be used immediately or stored by digital twin module 13420. Imported data may be collected from data streams, data batches, combinations thereof, etc. in response to a triggering event. The input module may receive data in a format suitable for communicating, reading, and/or writing information within digital twin module 13420.

In embodiments, the output module may transmit data to other system components (e.g., digital twin data store 13316, digital twin simulation system 13310, intelligence services 13004, etc.), devices, and/or client applications 13324. You can print or export. Data may be output as data streams, data bundles, combinations thereof, etc. in response to a triggering event (e.g., a request). The output module may output data in a format suitable for use or storage by the target element (e.g., one protocol for output to a client application and another protocol for digital twin data storage 13316).

In embodiments, an adapter module may process and/or convert data between an input module and an output module. In embodiments, an adapter module may transform and/or route data automatically (e.g., based on data type) or in response to a received request (e.g., in response to information within the data).

In embodiments, digital twin module 13420 may represent a set of delivery workpiece elements in a digital twin, and digital twin simulation system 13310 may represent a set of physical interactions of the workpiece elements with a worker or delivery robot. Simulate.

In embodiments, digital twin simulation system 13310 may determine process outcomes for simulated physical interactions that account for simulated human factors. For example, variations in workpiece throughput may cause, for example, operator response times to events, operator fatigue, discontinuities within operator actions (e.g., natural variations in human-movement speed, different positioning times). etc.), the effects of discontinuities on downstream processes, and the like. In embodiments, individualized worker interactions may be modeled using historical data collected, acquired, and/or stored by digital twin module 13420. The simulation may begin based on estimated amounts (e.g., worker age, industry average, workplace expectations, etc.). Simulations can also individualize data for each worker (e.g., comparing estimated quantities to collected worker-specific results).

In embodiments, information about workers (e.g., fatigue rates, efficiency rates, etc.) may be determined by analyzing the performance of specific workers over time and modeling that performance.

In an embodiment, digital twin module 13420 includes multiple proximity sensors within a sensor array. Proximity sensors may be configured or configured to detect elements of the environment or a delivery entity within a predetermined area. For example, proximity sensors may include electromagnetic sensors, optical sensors, and/or acoustic sensors.

Electromagnetic sensors may detect or be configured to detect objects or interactions through one or more electromagnetic fields (eg, emitted or received electromagnetic radiation). In embodiments, electromagnetic sensors include inductive sensors (e.g., radio frequency identification sensors), capacitive sensors (e.g., contact and non-contact capacitive sensors), combinations thereof, etc. .

Optical sensors may be configured or configured to detect objects or interactions via electromagnetic radiation, for example, in the far-infrared, near-infrared, optical, and/or ultraviolet spectra. In embodiments, optical sensors include image sensors (e.g., charge-coupled devices and CMOS active-pixel sensors), photoelectric sensors (e.g., transmitted-beam sensors, retroreflective sensors, and diffusion sensors), combinations thereof, etc. In embodiments, optical sensors may include liquid lens vision systems. Additionally, optical sensors may be implemented as part of a system or subsystem, such as a light detection and ranging (LIDAR) sensor.

Acoustic sensors may be configured or configured to detect objects or interactions through sound waves emitted and/or received by the acoustic sensors. In embodiments, acoustic sensors may include infrasound, acoustic, and/or ultrasonic sensors. Additionally, acoustic sensors may be grouped as part of a system or subsystem, such as a sound navigation and ranging (“SONAR”) sensor.

In embodiments, digital twin module 13420 stores and collects data from a set of proximity sensors. The collected data may be stored in digital twin data store 13316 for use by components of digital twin module 13420 and/or visualization by a user, for example. Such use and/or visualization may occur concurrently with the collection of the data or after the collection of the data (e.g., during later analysis and/or optimization of the process).

In embodiments, data collection may occur in response to a triggering condition. These triggering conditions may include, for example, expiration of a static or dynamic predetermined interval, obtaining a value below or above a static or dynamic value, receiving an automatically generated request or command from digital twin module 13420 or its components, interaction of elements with each sensor or sensors (e.g., in response to an object coming within a predetermined distance from a proximity sensor), user interaction with the digital twin (e.g., smart container digital twin, sensor array digital twin, or selection of a sensor digital twin), combinations thereof, etc.

In some embodiments, digital twin module 13420 collects and/or stores RFID data in response to the operator's or robot's interaction with real-world elements. For example, in response to a robot interaction with a smart container cargo, the digital twin will collect and/or store RFID data from RFID sensors associated with the corresponding cargo. Additionally or alternatively, robot interaction with the sensor-array digital twin will collect and/or store RFID data from RFID sensors within or associated with the corresponding sensor array. Similarly, robot interaction with a sensor digital twin will collect and/or store RFID data from the corresponding sensors. RFID data includes proximity RFID tags, RFID tag location, authorized RFID tags, unauthorized RFID tags, unrecognized RFID tags, RFID type (e.g., active or passive), error codes, and combinations thereof. It may include appropriate data achievable by RFID sensors such as .

In embodiments, digital twin module 13420 may further embed outputs from one or more devices within a corresponding digital twin. In embodiments, digital twin module 13420 embeds output from a set of individually-related devices into a delivery digital twin. For example, digital twin I/O system 13308 may receive information output from one or more wearable devices or mobile devices (not shown) associated with the individual. Wearable devices include image capture devices (e.g., body cameras or augmented reality headwear), navigation devices (e.g., GPS devices, inertial guidance systems), motion trackers, acoustic capture devices (e.g. For example, microphones), radiation detectors, combinations thereof, etc.

In embodiments, upon receiving output information, digital twin I/O system 13308 routes the information to digital twin configuration module 13318 to verify and/or update the delivery digital twin and/or associated digital twins. Additionally, digital twin module 13420 can use the embedded output to determine characteristics of the delivery entity or environment.

In an embodiment, digital twin module 13420 embeds output from a LIDAR point cloud system into a shipping digital twin. For example, the digital twin I/O system 13308 may receive information output from one or more LIDAR devices. LIDAR devices are configured to provide a plurality of points with associated location data (e.g., coordinates in absolute or relative x, y, and z values). Each of the plurality of points may include additional LIDAR properties such as intensity, number of returns, total returns, laser color data, return color data, scan angle, scan direction, etc. LIDAR devices may provide a point cloud containing a plurality of points to the digital twin module 13420, for example, via the digital twin I/O system 13308. Additionally or alternatively, digital twin module 13420 may receive a stream of points and assemble the stream into a point cloud, or may receive a point cloud and combine the received point cloud with existing point cloud data, map data, or It can be assembled with three-dimensional (3D) model data.

In embodiments, upon receiving output information, digital twin I/O system 13308 routes the point cloud information to digital twin configuration module 13318 to verify and/or update the delivery digital twin and/or associated digital twins. . In some embodiments, digital twin module 13420 is further configured to determine closed-shape objects in received LIDAR data. For example, the digital twin module 13420 groups a plurality of points within the point cloud as objects and, if necessary, blocks the objects' faces (e.g., the face of an object touching or adjacent to the floor or other objects such as other equipment). The surface of an object that touches or is adjacent to an object can be estimated. The system can use these closed-shaped objects to narrow the search space for digital twins, thereby increasing the efficiency of matching algorithms (e.g., shape matching algorithm).

In embodiments, digital twin module 13420 embeds output from a simultaneous location and mapping (“SLAM”) system in an environmental digital twin. For example, digital twin I/O system 13308 may receive information output from a SLAM system, such as a SLAM sensor, and embed the received information within an environmental digital twin corresponding to a location determined by the SLAM system. In embodiments, upon receiving output information from the SLAM system, digital twin I/O system 13308 routes the information to digital twin configuration module 13318 to verify and/or verify the delivery digital twin and/or associated digital twins. Update. This update automatically provides digital twins of unconnected elements without requiring user interaction with the digital twin module 13420.

In embodiments, digital twin module 13420 may utilize known digital twins to reduce computational requirements for a SLAM sensor by using suboptimal map-building algorithms. For example, suboptimal map-building algorithms can allow for higher uncertainty tolerance using simple boundary-area representations and identifying possible digital twins. Additionally or alternatively, digital twin module 13420 uses a boundary region representation to limit the number of digital twins, analyzes groups of potential twins to distinguish features, and then categorizes the individual digital twins. , perform a higher precision analysis on the distinguishing features to identify and/or remove categories of groups, or individual digital twins, and if a matching digital twin is not found, perform a fine scan of the remaining areas to be scanned. You can.

In embodiments, digital twin module 13420 utilizes data captured from other sensors in the environment (e.g., captured images or video, radio images, etc.) to perform an initial map-building process (e.g. , a simple boundary-area map or other suitable photogrammetric method), improve the simple boundary-area map by associating digital twins of known environmental objects with features of the simple boundary-area map, and further refine the remaining simple boundary-area maps. Further refinement of the map by performing precise scans can further reduce the computing required to construct the location map. In some embodiments, digital twin module 13420 can detect objects within the received mapping information and, for each detected object, determine whether the detected object corresponds to an existing digital twin of a real-world element. . In response to determining that a detected object does not correspond to an existing real-world element digital twin, digital twin module 13420 may create a new digital twin corresponding to the detected object using, for example, digital twin configuration module 13318. For example, you can create a detected object digital twin) and add the detected object digital twin to real-world element digital twins in the digital twin data store. Additionally or alternatively, in response to determining that a detected object corresponds to an existing real-world element digital twin, digital twin module 13420 may be configured to include new information detected by the simultaneous location and mapping sensor, if present. Digital twins of real-world elements can be updated.

In an embodiment, digital twin module 13420 displays the location of autonomously or remotely movable elements and their attributes within the delivery digital twin. These movable elements may include, for example, cargo, vehicles, autonomous vehicles, robots, etc. The positions of movable elements may be updated in response to a triggering condition. These triggering conditions may include, for example, expiration of a static or dynamic predetermined interval, receipt of an automatically generated request or command from digital twin module 13420 or its components, (e.g., an operator or machine destroys the beam), interaction of the respective sensor or element with the sensor (or in response to coming within a predetermined distance from a proximity sensor), interaction of the user with the digital twin (e.g., a delivery digital twin, a sensor array digital twin, or a sensor digital twin) selection), combinations thereof, etc.

In embodiments, time intervals may be based on the probability that each movable element has moved within a time period. For example, the time interval for updating robot position is relative to robots that are expected to move frequently (e.g., robots that perform tasks such as lifting and transporting cargo within and through container terminals). It can be relatively short for robots that are expected to move infrequently (for example, robots that perform the task of monitoring a process). Additionally or alternatively, the time interval may include increasing the time interval when movable elements are not detected, decreasing the time interval when or when the number of movable elements in the environment increases (e.g., increasing the number of robots and robot interactions), increasing the time interval during periods of reduced activity, decreasing the time interval during periods of unusual activity (e.g. inspections or maintenance). , reducing the time interval when unexpected or uncharacteristic movements are detected (e.g., frequent movements by typically static elements or coordinated movements, e.g., jointly to transport large objects). of moving robots), combinations thereof, etc. Additionally, the time interval may also include additional, semi-random acquisitions. For example, sometimes mid-interval positions may be acquired by digital twin module 13420 to enhance or evaluate the utility of a particular time interval.

In embodiments, digital twin module 13420 may analyze data received from digital twin I/O system 13308 to refine, remove, or add conditions. For example, the digital twin module 13420 may need to adjust the data collection time for movable elements that are updated more frequently than necessary (e.g., multiple consecutive received positions are the same or within a predetermined error margin). can be optimized.

In embodiments, digital twin module 13420 may receive, identify, and/or store a set of states associated with delivery entities or environments. A set of states may be, for example, data structures that include a plurality of properties and a set of identification criteria to uniquely identify each individual state. In embodiments, a set of states may cause digital twin module 13420 to set or change conditions of real-world elements and/or the environment (e.g., increase/decrease monitoring intervals, change operating conditions, etc.). ) can correspond to these desirable states.

In embodiments, the set of states may include, for example, the minimum monitored attributes for each state, a set of identification criteria for each state, and/or a set of conditions taken or recommended to be taken in response to each state. Additional available actions can be included. This information may be stored, for example, by digital twin data store 13316 or another data store. The set of states, or portions thereof, may be provided to, determined by, or changed by the digital twin module 13420. Additionally, the set of states may include data from separate sources. For example, details for identifying and/or responding to the occurrence of a first state may be provided to digital twin module 13420 via user input, and details for identifying and/or responding to the occurrence of a second state may be provided to the digital twin module 13420. Details may be provided to the digital twin module 13420 via an external system, and details may be provided to the digital twin module 13420 (e.g., through simulation or analysis of process data) to identify and/or respond to the occurrence of a third state. Details for identifying and/or responding to occurrences of the fourth state may be determined by module 13420 and stored by digital twin module 13420 as desired (e.g., responding to simulated occurrences of the state (or in response to analysis of data collected during the occurrence of a state and in response to the state).

In an embodiment, the plurality of attributes includes at least attributes necessary to identify each state. The plurality of attributes may be monitored or monitored when determining each state, but may further include additional attributes that do not necessarily identify each state. For example, the plurality of attributes for the first state may include related information such as rotational speed, battery level, energy input, linear speed, acceleration, temperature, strain, torque, volume, weight, etc.

The set of identification criteria may include information about each set of attributes to uniquely identify each state. Identification criteria may include, for example, rules, thresholds, limits, ranges, logical values, conditions, comparisons, combinations thereof, etc.

Changes in operating conditions or monitoring may be any suitable change. For example, after identifying the occurrence of each condition, digital twin module 13420 may increase or decrease the monitoring interval for the smart container without changing the operation of the smart container (e.g., between the nominal operation and Monitoring intervals can be reduced in response to different measured parameters). Additionally or alternatively, digital twin module 13420 may change the operation of the smart container (e.g., reduce speed or power input) without changing the monitoring of the smart container. In further embodiments, digital twin module 13420 may change the operation of the smart container (e.g., reduce speed or power input) and change the monitoring interval for the device (e.g., reduce the monitoring interval). Sikkim).

In embodiments, in accordance with some embodiments of the present disclosure, digital twin module 13420 may provide access by intelligent systems (e.g., intelligent service 13004) or users of digital twin module 13420. A set of identified sets of states associated with shipping entities and environments that can be identified and/or stored for the purpose of The set of states can be operational states (e.g., suboptimal, mediocre, optimal, critical, or alarm operation of one or more components), excess or deficiency states (e.g., supply-side or output-side quantities), these It may include combinations of .

In embodiments, digital twin module 13420 may monitor properties of real-world elements and/or digital twins to determine individual status. Attributes may be, for example, operating conditions, set points, threshold points, status indicators, other sensed information, combinations thereof, etc. For example, properties may include power input, operating speed, threshold speed, and operating temperature of the monitored elements. Although the illustrated example illustrates uniform monitored properties, the monitored properties may differ depending on the target device (e.g., digital twin module 13420 does not monitor rotational speed for objects that do not have rotatable components). won't).

Each set of states includes a set of identification criteria that meet certain criteria that are unique among the group of monitored sets of states. Digital twin module 13420 may identify an overspeed condition, for example, in response to monitored attributes meeting a first set of identification criteria (e.g., operating speed is greater than a threshold speed). .

In response to determining that one or more sets of conditions exist or have occurred, digital twin module 13420 updates triggering conditions for one or more monitoring protocols, issues an alert or notification, or sends a sub-set of digital twin module 13420. You can trigger actions of components. For example, subcomponents of digital twin module 13420 can take steps to mitigate and/or evaluate the effects of a set of detected conditions. When attempting to take actions to mitigate the effects of a detected set of conditions on real-world elements, digital twin module 13420 determines whether instructions exist (e.g., stored in digital twin data store 13316). ) or whether it should be developed (e.g., through simulation and intelligence services or through user or operator input). Additionally, digital twin module 13420 may be configured to configure a set of detected conditions, for example, concurrently with mitigation actions or in response to digital twin module 13420 determining that it does not have stored mitigation instructions for a set of detected conditions. The effects of the set can be evaluated.

In an embodiment, digital twin module 13420 uses digital twin simulation system 13310 to simulate one or more impacts, such as immediate, upstream, downstream, and/or ongoing effects of recognized conditions. The digital twin simulation system 13310 may collect and/or be provided with values associated with the set of evaluated states. In simulating the impact of one or more sets of states, digital twin simulation system 13310 may recursively evaluate the performance characteristics of the affected digital twins until convergence is achieved. Digital twin simulation system 13310 may, for example, operate in conjunction with intelligence service 13004 to determine response actions to mitigate, mitigate, contain, and/or prevent the occurrence of one or more sets of conditions. For example, digital twin simulation system 13310 may recursively simulate the effects of one or more sets of states until a desired fit is achieved (e.g., until convergence is achieved) and determine potential actions. may provide simulated values to intelligence service 13004 for evaluation and decision, receive potential actions, and/or evaluate the impact of each of the potential actions on each desired fit (e.g. , cost functions to minimize production failures, preserve critical components, minimize maintenance and/or downtime, optimize system, operator, user, or personal safety, etc.).

In embodiments, digital twin simulation system 13310 and intelligence service 13004 may perform each evaluation until desired conditions are met (e.g., convergence for each evaluated cost function for each evaluated action). Simulated values and response actions for desired results can be shared and updated repeatedly. Digital twin module 13420 may store the results in digital twin data store 13316 for use in response to determining that one or more sets of conditions have occurred. Additionally, simulation and evaluation by digital twin simulation system 13310 and/or intelligence service 13004 may occur in response to the occurrence or detection of an event.

In embodiments, simulations and evaluations are triggered only when associated actions do not exist within digital twin module 13420. In further embodiments, simulations and evaluations can be performed concurrently with the use of stored actions to evaluate the utility or effectiveness of the actions in real time and/or determine whether additional actions should be used or whether unrecognized conditions may have occurred. Evaluate. In embodiments, intelligence service 13004 may also be provided with notifications of instances of undesirable actions with or without data about the undesirable aspects or consequences of such actions to optimize later evaluations. You can.

In an embodiment, digital twin module 13420 evaluates and/or displays the impact of downtime of a smart container within a digital twin of a smart container fleet. For example, digital twin module 13420 may use digital twin simulation system 13310 to simulate the immediate, upstream, downstream, and/or ongoing effects of a smart container downtime condition. Digital twin simulation system 13310 determines optimal, suboptimal, and minimum performance requirements for elements within affected digital twins (e.g., real-world elements and/or nested digital twins), and /or collect performance-related values such as their characteristics available to the affected digital twins, impact on nested digital twins, redundant systems within the affected digital twins, combinations thereof, etc. It can be provided.

In embodiments, digital twin module 13420 may: use real-world element digital twins to simulate one or more operating parameters for real-world elements in response to a shipping entity or environment provided with given characteristics; calculate a mitigation action to be taken by one or more of the real-world elements in response to being provided with concurrent characteristics; In response to detecting concurrent characteristics, the device is configured to actuate mitigation action. The calculation may be performed in response to detecting simultaneous characteristics or operating parameters that are external to the respective design parameters, or may be determined through simulation prior to detection of such characteristics.

Additionally or alternatively, digital twin module 13420 may provide alerts to one or more users or system elements in response to detecting a condition.

In an embodiment, digital twin I/O system 13308 includes a routing module. The routing module collects navigation data from the elements and transmits the navigation data to components of the digital twin module 13420 (e.g., digital twin simulation system 13310, digital twin dynamic model system 13312, and/or Intelligent service 13004) may provide and/or request navigation data and output navigation data to elements (e.g., wearable devices). Navigation data may be collected or estimated using, for example, historical data, guidance data provided in elements, combinations thereof, etc.

For example, navigation data may be collected or estimated using historical data stored by digital twin module 13420. Historical data includes acquisition time, relevant factors, polling intervals, tasks performed, loaded or unloaded conditions, whether previous guidance data was provided and/or followed, conditions of the delivery entity or environment, and other elements within the delivery entity or environment. , combinations thereof, etc. The estimated data may be determined using one or more suitable routing algorithms. For example, the estimated data may be calculated using an appropriate order-picking algorithm, an appropriate path search algorithm, a combination thereof, etc. The order selection algorithm may be, for example, a maximum gap algorithm, an s-shape algorithm, an aisle-by-aisle algorithm, a combined algorithm, combinations thereof, etc. The path search algorithm may be, for example, Dijkstra's algorithm, A* algorithm, hierarchical path search algorithm, incremental path search algorithm, arbitrary angular path search algorithm, flow field algorithm, combinations thereof, etc.

In embodiments, digital twin module 13420 collects navigation data for a set of smart containers for representation in the digital twin. Additionally or alternatively, digital twin module 13420 collects navigation data for a set of mobile equipment assets in a delivery environment into a digital twin.

In an embodiment, digital twin module 13420 collects a system for modeling traffic of mobile elements (e.g., smart containers, container ships, robots, trucks, trains, cargo, etc.) in a shipping digital twin. For example, digital twin module 13420 can model traffic patterns for a set of smart containers, mobile equipment assets, cargo combinations thereof, etc. Traffic patterns can be estimated based on modeling traffic patterns from historical data and concurrently collected data. Additionally, traffic patterns can be updated continuously or intermittently depending on conditions.

Digital twin module 13420 may change traffic patterns (e.g., by providing updated navigation data to one or more of the mobile elements) to achieve one or more predetermined criteria. Predetermined criteria include, for example, increasing process efficiency, reducing interactions between smart containers and mobile device assets, minimizing smart container path length, people's paths or potential paths. This may include routing smart containers around, combinations of these, etc.

In embodiments, digital twin module 13420 may provide traffic data and/or navigation information to mobile elements within a delivery digital twin. Navigation information may be provided as commands or rule sets, displayed route data, or selective operation of devices. For example, digital twin module 13420 may provide a set of commands to a smart container to guide the smart container from a source location to one or more designated locations along a route and/or along a desired path. A smart container can communicate updates to the system, including obstacles, reroutes, and unexpected interactions with other assets along the path.

In some embodiments, an ant-based system allows shipping entities, including smart containers, to send other shipping containers and/or shipping entities, including themselves, with one or more messages to follow on later journeys. Allows you to leave a trail. In embodiments, the messages include information related to measurement collection.

In embodiments, digital twin module 13420 includes design specification information for representing real-world elements using a digital twin. A digital twin can correspond to existing or potential real-world elements. Design specification information may be received from one or more sources. For example, design specification information may be set by user input, determined by digital twin module 13420 (e.g., via digital twin simulation system 13310), and may be determined by the user or digital twin simulation system 13310. It may include design parameters optimized by , combinations thereof, etc. Digital twin simulation system 13310 may present design specification information for a component to a user, for example, through a monitor or virtual reality headset. Design specification information may be displayed schematically (eg, as part of a process diagram or table of information) or as part of an augmented reality or virtual reality display. Design specification information may be displayed, for example, in response to user interaction with digital twin module 13420 (e.g., generally through user selection or user selection of elements to include design specification information within the display). It can be. Additionally or alternatively, design specification information may be automatically displayed, for example, when an element comes into view of an augmented reality or virtual reality device. In embodiments, the displayed design specification information may be an indication of an information source (e.g., different displayed colors indicate user input versus digital twin module 13420 decisions), (e.g., between design specification information and operational information). ) may further include indications of mismatch, combinations thereof, etc. In some embodiments, digital twin module 13420 may provide an augmented reality view that displays to the wearer discrepancies between design parameters or expected parameters of real-world elements. The displayed information may correspond to real-world elements that are not within the wearer's field of view (eg, elements in another room or obscured by a machine). This allows operators to quickly and accurately resolve discrepancies to determine one or more sources of discrepancies. The cause of the discrepancy can then be determined, for example, by digital twin module 13420 and ordered corrective actions. In example embodiments, the wearer can view malfunctioning subcomponents of the machines without removing occluding elements (eg, housings or shields). Additionally or alternatively, the wearer may provide, for example, a display of the removal process (e.g., location of fasteners to be removed), assemblies or subassemblies that must be transported to other areas for repair (e.g., , dust-sensitive components), assemblies or subassemblies that require lubrication, and the locations of objects for reassembly (e.g., stores where the wearer placed removed objects and facilitates reassembly) Instructions for repairing the device may be provided, including directing the wearer or another wearer to stored locations to minimize further disassembly or missing parts from the reassembled elements. This expedites repair operations, minimizes process impact and allows operators to disassemble and reassemble equipment (e.g. by coordinating disassembly without direct communication between operators) and return it to service (e.g. Equipment life and reliability can be increased (by ensuring that all components are properly replaced before deployment), combinations of these, and more.

In embodiments, digital twin module 13420 may include, integrate, integrate with, manage, handle, link, provide input, output, control, coordinate, or otherwise interact with digital twin dynamic model system 13312. there is. Digital Twin Dynamic Model System 13312 is a physical transportation asset, workers, processes, transportation facility, warehouse, etc., in such a way that a digital twin can represent these shipping entities and environments, and their properties or characteristics, in real time or very near real time. Update the properties of a digital twin of a set of delivery entities and/or environments, including the properties of (or any of the other types of entities or environments described in this disclosure or documents incorporated herein by reference). there is. In some embodiments, digital twin dynamic model system 13312 may acquire sensor data received from a sensor system and determine one or more attributes of a delivery environment or delivery entity based on the sensor data and based on one or more dynamic models. You can decide.

In an embodiment, the digital twin dynamic model system 13312 may include vibration values, failure probability values, downtime probability values, downtime cost values, price values, energy values, performance values, financial values, temperature values, and humidity values. , heat flow value, cargo loading value, fluid flow value, radiation value, material concentration value, velocity value, acceleration value, position value, pressure value, stress value, strain value, light intensity value, sound level value, volume value, shape. The values of various properties of the digital twin and/or one or more embedded digital twins may be updated/assigned, including but not limited to properties, material properties, and dimensions.

In embodiments, a digital twin may be comprised of other embedded digital twins (e.g., via references). For example, a digital twin of a container terminal may include an embedded digital twin of a container ship and one or more embedded digital twins of each of one or more smart containers enclosed within the container ship. The digital twin may be embedded, for example, in the memory of a smart container with onboard IT systems (e.g., memory of onboard diagnostic systems, control systems (e.g., SCADA systems), etc.). Other non-limiting examples of where a digital twin can be embedded include: on a worker's wearable device; Within memory on local network assets such as switches, routers, access points, etc.; Within cloud computing resources provisioned for an environment or entity; and on asset tags or other memory structures dedicated to the entity.

In embodiments, digital twin dynamic model system 13312 may update properties of a digital twin and/or one or more embedded digital twins on behalf of a client application 13324 (such as smart container system 13000). In embodiments, client application 13324 may be an application related to smart container system 13000, a shipping component, or the environment (e.g., monitoring a shipping facility or components within it, simulating a shipping environment, etc.). In embodiments, client application 13324 may be used in connection with both fixed and mobile data collection systems. In embodiments, client application 13324 may be used in connection with an Industrial Internet of Things sensor system.

In embodiments, digital twin dynamic model system 13312 utilizes digital twin dynamic models 13374 to model the behavior of the delivery entity and/or environment. Dynamic models 13374 use a limited number of measurements to enrich the digital representation of the delivery entity and/or environment, such as based on scientific principles, so that digital twins include interactions of the delivery entities. It can make it possible to express physical reality. In embodiments, dynamic models 13374 are formal or mathematical models. In embodiments, dynamic models 13374 may be based on scientific laws, laws of nature, and formulas (e.g., Newton's laws of motion, second law of thermodynamics, Bernoulli's principle, ideal gas law, Dalton's law of partial pressures). , Hooke's law of elasticity, Fourier's law of heat conduction, Archimedes' principle of buoyancy, etc.). In embodiments, dynamic models are machine learning models.

In an embodiment, digital twin module 13420 may have a digital twin dynamic model data store 13376 for storing dynamic models 13374 that can be represented as digital twins. In embodiments, digital twin dynamic model data store 13376 may be searchable and/or searchable. In embodiments, digital twin dynamic model data store 13376 includes metadata that allows a user to understand what properties a given dynamic model can handle, what inputs are required, what outputs are provided, etc. can do. In some embodiments, digital twin dynamic model data store 13376 may allow models to be further deepened or further simplified, for example, based on the range of data and/or inputs available, the granularity of the inputs, and/or situational factors. It may be hierarchical in cases where it is feasible (e.g., when something becomes of high interest and higher fidelity models are accessed over a period of time).

In embodiments, a digital twin or digital representation of a delivery entity or environment collectively defines a set of attributes of the represented physical delivery assets, devices, personnel, processes, facilities, and/or environment, and/or their possible behaviors. It may contain a set of data structures. In embodiments, digital twin dynamic model system 13312 may utilize dynamic models 13374 to inform a set of data structures that collectively define a digital twin with real-time data values. Digital twin dynamic models 13374 receive one or more sensor measurements, industrial Internet of Things device data, and/or other suitable data as inputs, and create one or more dynamic models 13374 based on the received data and one or more dynamic models 13374. Outputs can be calculated. The digital twin dynamic model system 13312 then uses one or more outputs to update the digital twin data structure.

In one example, the set of properties of the digital twin of a shipping entity that can be updated by the digital twin dynamic model system 13312 using dynamic models 13374 include vibration characteristics of the shipping entity, temperature(s) of the shipping entity, and ), state of the delivery entity (e.g., solid, liquid, or gas), location of the delivery entity, displacement of the delivery entity, velocity of the delivery entity, acceleration of the delivery entity, probability of downtime values associated with the delivery entity, delivery Cost of downtime values associated with the entity, financial information associated with the delivery entity, heat flow characteristics associated with the delivery entity, fluid flow rates associated with the delivery entity (e.g., fluid flow rates of a fluid flowing through a pipe), It may include identifiers of other digital twins embedded within the digital twin of the shipping entity and/or identifiers of digital twins that embed the digital twin of the shipping entity, and/or other suitable properties. Dynamic models 13374 associated with an asset's digital twin calculate values for such asset digital twin properties based on input data collected from sensors and/or devices deployed in the industrial environment and/or other appropriate data. , interpolate, extrapolate, and/or output, and subsequently populate the asset digital twin with the calculated values.

In some embodiments, the set of properties of the digital twin of the delivery device that can be updated by the digital twin dynamic model system 13312 using dynamic models 13374 include the state of the device, the location of the device, and the temperature of the device. (s), the trajectory of the device, the identifiers of other digital twins with which the device's digital twin is embedded, embedded, linked, included, integrated, takes input, provides output and/or interacts, etc. can do. Dynamic models 13374 associated with a device's digital twin may be configured to calculate or output values for these device digital twin properties based on input data and subsequently update the device digital twin with the calculated values.

Exemplary properties of the digital twin of the delivery environment that can be updated by the digital twin dynamic model system 13312 using dynamic models 13374 include dimensions of the delivery environment, temperature(s) of the delivery environment, and Humidity value(s), fluid flow characteristics in the delivery environment, heat flow characteristics of the delivery environment, lighting characteristics of the delivery environment, acoustic characteristics of the delivery environment, physical objects in the environment, processes occurring in the delivery environment. , flows of the delivery environment (in the case of a body of water), etc. Dynamic models associated with the digital twin of the delivery environment calculate or output these properties based on input data and/or other appropriate data collected from sensors and/or devices deployed in the delivery environment, and subsequently calculate the calculated values. It can be configured to populate the delivery environment digital twin.

In embodiments, dynamic models 13374 may adhere to physical constraints that define boundary conditions, constants, or variables for digital twin modeling. For example, the physical characterization of a shipping entity or a digital twin of a shipping environment may include the gravitational constant (e.g. 9.8 m/s ² ), the coefficient of friction of the surface, the thermal coefficient of the material, the maximum temperature of the asset, the maximum flow capacity, etc. It can be included. Additionally or alternatively, dynamic models may follow natural laws. For example, a dynamic model may adhere to the laws of thermodynamics, laws of motion, laws of fluid mechanics, laws of buoyancy, laws of heat transfer, laws of radiation, laws of quantum mechanics, etc. In some embodiments, the dynamic model may adhere to biological aging theory or mechanical aging principles. Accordingly, when the digital twin dynamic model system 13312 facilitates real-time digital representations, the digital representations can conform to the dynamic models so that the digital representations mimic real-world conditions. In some embodiments, the output(s) from the dynamic model may be presented to a human user and/or compared against real-world data to ensure convergence of the dynamic models with the real world. Additionally, because dynamic models are based in part on assumptions, the characteristics of the digital twin can be improved and/or corrected when real-world behavior differs from the digital twin's behavior. In embodiments, it is recognized that input is missing from the desired dynamic model, the model in operation is not behaving as expected (perhaps due to missing and/or defective sensor information), and a different result is needed (e.g., a high-interest Additional data collection and/or measurement may be recommended, based on (due to situational factors that make it present), etc.

Dynamic models can be obtained from a number of different sources. In some embodiments, users may upload models created by the user or a third party. Additionally or alternatively, models may be created on a digital twin system using a graphical user interface. Dynamic models may include custom models constructed for a specific environment and/or set of delivery entities and/or agnostic models applicable to digital twins of similar types. Dynamic models may be machine learning models. In an embodiment, the dynamic model may be a machine learning model provided by intelligence service 13004.

In embodiments, the digital twin dynamic model system 13312 may be used to enable collected sensor data from sensor systems, data collected from Internet of Things connected devices 13338, and/or delivery digital twins. One or more dynamic models 13374 to update a set of properties of the digital twin and/or one or more embedded digital twins on behalf of the client application 13324 based on the influence of other appropriate data within the set of dynamic models 13374 ) is used. In embodiments, digital twin dynamic model system 13312 uses one or more digital twins to represent physical delivery entities, devices, workers, processes, and/or delivery environments that are managed, maintained, and/or monitored by client application 13324. can be instructed to run a specific dynamic model.

In embodiments, the digital twin dynamic model system 13312 may obtain data from other types of external data sources, not necessarily shipping data sources, but may provide data that can be used as input data to the dynamic model. For example, traffic data, truck fleet data, aviation data, road data, freight data, maritime data, weather data, news events, social media data, etc. can be generated by sensor data, Industrial Internet of Things device data, and/or dynamic models. It can be collected, crawled and requested to supplement other data used. In embodiments, digital twin dynamic model system 13312 may obtain data from machine vision module 13422. Machine vision module 13422 may use video and/or still images to provide measurements (e.g., positions, states, etc.) that can be used as inputs by dynamic models.

In embodiments, digital twin dynamic model system 13312 may feed this data to one or more of the dynamic models discussed above to obtain one or more outputs. These outputs include calculated vibration characteristics, failure probability values, downtime probability values, downtime cost values, downtime values, temperature values, pressure values, humidity values, precipitation values, visibility values, air quality values, and strain values. , stress values, displacement values, velocity values, acceleration values, position values, performance values, financial values, price values, electrodynamic values, thermodynamic values, fluid flow values, etc. Client application 13324 can then initiate a digital twin visualization event using the results obtained by digital twin dynamic model system 13312. In embodiments, the visualization may be a heat map visualization.

As illustrated by FIG. 148, digital twin dynamic model system 13312 updates one or more properties of digital twins of shipping entities and/or environments such that the digital twins represent the shipping entities and/or environments in real time. Requests can be received. At step A100, digital twin dynamic model system 13312 receives a request to update one or more properties of the digital twin of the shipping entity and/or environment. For example, digital twin dynamic model system 13312 may receive a request from a client application 13324 or from another process executed by digital twin module 13420 (e.g., a predictive maintenance process). A request may indicate one or more properties and a digital twin or digital twins implicated by the request. At step A102, digital twin dynamic model system 13312 determines one or more digital twins required to fulfill the request and retrieves one or more required digital twins, including any embedded digital twins, from digital twin data repository 13316. Search. At step A104, digital twin dynamic model system 13312 determines one or more dynamic models needed to fulfill the request and retrieves the one or more required dynamic models from digital twin dynamic model data repository 13376. At step A106, the digital twin dynamic model system 13312 connects one or more sensors, Internet of Things connected devices 13338 from the sensor system based on the available data sources and one or more required inputs of the dynamic model(s). ), and/or other data sources from the digital twin I/O system 13308. In embodiments, data sources may be defined from inputs required by one or more dynamic models or may be selected using a lookup table. At step A108, digital twin dynamic model system 13312 retrieves selected data from digital twin I/O system 13308. At step A110, the digital twin dynamic model system 13312 executes the dynamic model(s) using the retrieved input data (e.g., speed sensor data, image data, etc.) as input, and executes the dynamic model(s) and the input Determines one or more output values based on the data. At step A112, digital twin dynamic model system 13312 updates the value of one or more properties of one or more digital twins based on one or more outputs of the dynamic model(s).

In an example embodiment, client application 13324 may be configured to provide a digital representation and/or visualization of a digital twin of a shipping entity. In embodiments, client application 13324 may include one or more software modules executed by one or more server devices. These software modules may be configured to quantify properties of the digital twin, model properties of the digital twin, and/or visualize digital twin behaviors. In embodiments, these software modules may allow users to select specific digital twin behavior visualizations for viewing. In embodiments, these software modules may allow a user to choose to view a digital twin behavior visualization playback. In some embodiments, client application 13324 may provide selected behavior visualizations to digital twin dynamic model system 13312.

In embodiments, digital twin dynamic model system 13312 may receive a request from client application 13324 to update properties of the digital twin to enable a digital representation of the delivery entity and/or environment, wherein real-time A digital representation is a visualization of the digital twin. In embodiments, a digital twin may be rendered by a computing device so that a human user can view digital representations of real-world delivery entities, devices, workers, processes, and/or environments. For example, a digital twin can provide results to a display device. In embodiments, dynamic model outputs and/or related data may be overlaid on a rendering of the digital twin. In embodiments, dynamic model outputs and/or related information may appear along with a rendering of the digital twin in a display interface. In embodiments, the relevant information may include real-time video footage associated with the real-world entity represented by the digital twin. In embodiments, the relevant information may be graphical information. In embodiments, the graphical information may depict motion, where a user is enabled to select a view of the graphical information in the x, y, and z dimensions. In embodiments, the relevant information may be cost data including daily downtime cost, repair data cost, new part data cost, new machine data cost, etc. In embodiments, the relevant information may be probability of downtime data, probability of failure data, etc. In embodiments, the relevant information may be time to failure data.

In embodiments, relevant information may be recommendations and/or insights. For example, recommendations or insights received from an intelligence service related to a smart container may be displayed along with a rendering of the smart container's digital twin in a display interface.

In embodiments, clicking, touching, or otherwise interacting with a rendered digital twin in a display interface allows the user to “drill down” and view the underlying subsystems or processes and/or embedded digital twins. It can be done. In embodiments, clicking, touching, or otherwise interacting with information related to a digital twin rendered in a display interface may allow a user to “drill down” and view basic information.

In some embodiments, the digital twin dynamic model system 13312 may receive a request from a client application 13324 to update properties of the digital twin to enable a digital representation of the delivery entity and/or environment, where the digital twin The representation is a heat map visualization of the digital twin. In embodiments, a platform is provided with heat maps that display data outputs from dynamic models 13374, Internet of Things connected devices 13338, and data collected from sensor systems to provide input to a display interface. do. In embodiments, the heat map interface may be configured to handle and provide information for visualization of various sensor data, dynamic model output data, and other data (e.g., map data, analog sensor data, and other data). As an output for digital twin data, it is provided to other systems, such as mobile devices, tablets, dashboards, computers, AR/VR devices, etc. The digital twin representation is a form factor suitable for conveying visual input to the user, such as the presentation of a map containing indicators of levels of analog sensor data, digital sensor data, and output values from dynamic models (e.g. It may be provided as a user device, VR-enabled device, AR-enabled device, etc.). In embodiments, signals from various sensors or input sources (or optional combinations, permutations, mixtures, etc.) as well as data determined by digital twin dynamic model system 13312 can be used to store input data into a heat map. can be provided. Coordinates may be real-world location coordinates (such as geographic location or location on a map), as well as other coordinates such as time-based coordinates, frequency-based coordinates, or analog sensor signals, digital sensor signals, dynamic model outputs, Input source information, and other coordinates allowing the representation of various combinations, and/or colors, may be included in a map-based visualization to represent various levels of input along relevant dimensions. For example, if a container terminal is operating at a critical level condition (e.g. due to excessive traffic or delays), among many other possibilities, the heat map interface could alert the user by marking the container port orange. You can. In the example of a heat map, clicking, touching, or otherwise interacting with the heat map allows the user to drill down into underlying container vessels, dynamic model outputs, or other inputs that are used as input to the heat map display. Make data visible. In other examples, such as when a digital twin is displayed in a VR or AR environment, where a smart container mechanical component is vibrating outside of normal operation, a haptic interface may induce vibration when the user touches a representation of the mechanical component, or If a machine component is operating in an unsafe manner, directional sound signals can direct the user's attention toward the machine in the digital twin, such as playing from specific speakers in a headset or other sound system.

In embodiments, the digital twin dynamic model system 13312 may take a set of ambient environmental data and/or other data and/or environmental data within a set of dynamic models 13374 that are used to enable the digital twin. Alternatively, a set of attributes of a digital twin of a shipping entity or environment can be automatically updated based on the influence of other data. Surrounding environment data may include temperature data, barometric pressure data, humidity data, wind data, rainfall data, tide data, storm surge data, cloud cover data, current data, snowfall data, visibility data, water level data, etc. . Additionally or alternatively, the digital twin dynamic model system 13312 may be used to enable a digital twin of Internet of Things connected devices 13338 deployed in an industrial environment as inputs to a set of dynamic models 13374 used to enable the digital twin. A set of environmental data measurements collected by set can be used. For example, the digital twin dynamic model system 13312 may be used to describe, for example, machines, devices, components, parts, operations, functions, conditions, states, events, workflows, and other delivery environments. Cameras, monitors, embedded sensors, mobile devices, diagnostic devices and systems, metrology systems, for monitoring various parameters and characteristics of elements (collectively covered by the term “states”) Dynamic models 13374 may be fed with data collected, handled, or exchanged by Internet of Things connected devices, such as telematics systems. Other examples of Internet of Things connected devices include smart fire alarms, smart security systems, smart air quality monitors, smart/learning thermostats, and smart lighting systems.

Figure 149 shows an example embodiment of a method for updating a set of downtime cost values in a digital twin of a smart container. In this example, digital twin dynamic model system 13312 may receive requests from client application 13324 to populate the smart container fleet digital twin with downtime real-time cost values associated with a smart container. At step B200, the digital twin dynamic model system 13312 receives one or more downtime cost values of the smart container digital twin 13504 and any embedded digital twins (e.g., robots, cargo, etc.) from the client application 13324. A request to update is received from the client application 13324. Next, at step B202, the digital twin dynamic model system 13312 determines one or more digital twins needed to fulfill the request and retrieves the one or more required digital twins. In this example, digital twin dynamic model system 13312 can retrieve the fleet's digital twins, smart containers, and any other embedded digital twins from digital twin data store 13316. At step B204, digital twin dynamic model system 13312 determines one or more dynamic models needed to fulfill the request and retrieves the one or more required dynamic models from dynamic model data store 13316. At step B206, digital twin dynamic model system 13312 generates a dynamic model via available data sources (e.g., available sensors from a set of sensors within a sensor system) and digital twin I/O system 13308. Select dynamic model input data sources (e.g., one or more sensors from a sensor system and/or any other suitable data) based on one or more desired inputs. In this example, the retrieved dynamic model(s) may be configured to take historical downtime data and operational data as input and output data representing the cost of daily downtime for smart containers in a smart container fleet. At step B208, digital twin dynamic model system 13312 retrieves historical downtime data and operational data from digital twin I/O system 13308. At step B210, digital twin dynamic model system 13312 executes dynamic model(s) using the retrieved data as input and calculates one or more outputs representing the cost of daily downtime for smart containers within a smart container fleet. . Next, at step B212, digital twin dynamic model system 13312 updates one or more downtime cost values of smart container digital twin 13504 and embedded digital twins based on one or more outputs of the dynamic model(s).

In embodiments, smart container system 13000 includes a robotic process automation (RPA) module 13416 configured to automate internal shipping workflows based on robotic process automation. The RPA module 13416 can develop program interfaces for user interfaces of external systems such as devices, programs, networks, databases, etc. The RPA module 13416 is configured to enable the smart container system 13000 to interface with an external system without using an application programming interface (API) or in addition to an API. The RPA module 13416 can develop an action list by watching the user perform a task in a graphical user interface (GUI) and recording the task in the action list. The RPA module 13416 can automate the workflow by repeating tasks in the action list within the GUI.

In some embodiments, RPA module 13416 may include and/or communicate with an intelligence service 13004 configured to perform robotic process automation processes. Intelligence service 13004 may use one or more machine learning techniques to develop one or more machine learning models. Machine learning models may develop, define, and/or implement RPA-based program interfaces to facilitate interfacing of system 13000 with one or more external devices.

RPA module 13416 may be needed for smart container system 13000 to communicate with external systems that do not have an API or have an outdated API. For example, RPA module 13416 may allow smart container system 13000 to interface with older external devices that do not contain an API or have an outdated API. RPA module 13416 may allow smart container system 13000 to interface with external systems similar to how a user interfaces with external systems, such as through the external system's user interface. In some embodiments, RPA module 13416 allows smart container system 13000 to emulate actions and/or sequences of actions that can be performed by a user to interface with external systems. Examples of programmatic interfacing by the RPA module 13416 to interface with an external system include manipulating a markup language, emulating computer mouse movements, and/or “clicking” on one or more elements of the user interface. , entering information into fillable fields and submitting the information through a client program and/or portal, and transmitting digital signals that appear to be transmitted from the user device to an external system.

In some embodiments, RPA module 13416 may be configured to facilitate communication with new and/or updated external systems. When a new external system is developed or an external system is updated, the RPA module 13416 is configured to update the New and/or updated program interfaces may be developed by container system 13000 to facilitate interfacing with new and/or updated external systems. For example, RPA module 13416 may enable smart container system 13000 to interface with new and/or updated external devices in a manner consistent with how smart container system 13000 interfaced with older external devices. , may be configured to provide input to an old external device, provide input to a new and/or updated external device, compare related outputs, and adjust the input to the new and/or updated external device.

In some embodiments, RPA module 13416 may act as an API to legacy and/or external systems. The RPA module 13416 is configured to externally represent the smart container system 13000 as having an API capable of interfacing with one or more external devices or otherwise programmatically handling signals transmitted by external devices. may be, where the RPA module 13416 has developed a programmatic interface to handle such requests other than the API. For example, an old external system could be configured to communicate through a series of signals understood by an old API. The RPA module 13416 may configure the smart container system 13000 to operate as if the smart container system 13000 included an outdated API.

In some embodiments, RPA module 13416 may be configured to provide a user interface for use by one or more users of smart container system 13000. Intelligence service 13004 may generate, by one or more machine learning methods, a user interface that allows a user to interface with one or more components and/or functionality of smart container system 13000. RPA module 13416 may use robotic process automation techniques to operate a user interface created by intelligence service 13004. Intelligent services 13004 may interface with the user interface based on variables such as changes in demand conditions, new and/or modified functionality of the smart container system 13000, new and/or modified conditions in systems external to the smart container system 13000, etc. can be dynamically created and/or adjusted. Examples of new and/or modified conditions of systems external to the smart container system 13000 may include changes to third-party service offerings, regulatory changes, etc.

In some embodiments, RPA module 13416 may be configured to avoid detection of robotic process automation by systems external to smart container system 13000. Some of the external systems may be designed to attempt to detect when an external system is communicating with a system that uses robotic process automation, such as smart container system 13000. Upon detecting that the smart container system 13000 is using robotic process automation, the external system may limit, remove, or modify the ability of the smart container system 13000 to communicate with the external system. The RPA module 13416 may “trick” external systems into believing that the RPA module 13416 is a human user to avoid detection of robotic process automation and to avoid limiting or eliminating communications by external systems. It can emulate human interfacing with external systems. RPA module 13416 may, for example, dynamically change paths of interaction with external systems, interact with user interface elements with inconsistent timing, and avoid “misclicks” or “typos.” Detection can be avoided by making human-like errors such as "typos" etc.

In some embodiments, intelligence service 13004 may be configured to create a machine learning model to avoid detection of robotic process automation. A machine learning model can be created by using data from interactions with one or more graphical user interfaces by real humans and developing robotic process automation techniques that emulate the ways in which real humans interface with one or more graphical user interfaces. . For example, training data may include mouse and/or touch timing and accuracy, typing speed and accuracy, elements of the graphical user interface used, etc.

In some embodiments, RPA module 13416 may be configured to verify data transmitted to and/or received from external systems. The RPA module 13416 receives data sent to the smart container system 13000 by users of the external system, data sent to the smart container system 13000 by users of the smart container system 13000, and/or Users of container system 13000 may verify one or more of the data transmitted to an external system. The RPA module 13416 performs optical character recognition, performs image recognition and/or processing, identifies data stored in web pages, receives data from a back-end database of an external system, and receives data from a back-end database of the smart container system 13000. Data may be verified by one or more of data reception, etc.

In some embodiments, intelligence services 13004 may be configured to develop one or more machine learning models for data validation. For example, intelligence service 13004 may train data transmitted by users and/or received from one or more databases and/or sources external to smart container system 13000 to “learn” to identify valid data. It can be used as data. Intelligence service 13004 may send one or more machine learning models to RPA module 13416 for data validation. RPA module 13416 may implement one or more machine learning models for data verification.

In some embodiments, RPA module 13416 may be configured to facilitate verification of processes performed by RPA module 13416. The RPA module 13416 generates a plurality of process validation logs when the RPA module 13416 performs one or more processes related to the smart container system 13000 and/or external systems on behalf of one or more users. can do. Process verification logs may include one or more of timestamps, transaction receipts, user interface screenshots, or any other suitable data entry, file, etc., to provide verification of the processes performed by the RPA module 13416. You can. The RPA module 13416 may store the process verification log in one or more databases and may transmit the process verification log to the smart container system 13000 and/or a user of the smart container system 13000. The RPA module 13416 may automatically transmit a process verification log according to a schedule, upon request by a user of the smart container system 13000, when one or more conditions are true, etc.

In some embodiments, RPA module 13416 may be configured to adjust the behavior of the robotic process automation in response to feedback obtained through one or both of data validation and process validation. A user of smart container system 13000 may view validation of data provided by RPA module 13416 and, in response to validation of the data, command RPA module 13416 to adjust the behavior of RPA module 13416. You can. A user of smart container system 13000 may view one or more of the process verification logs and, in response to one or more process verification logs, command RPA module 13416 to adjust the behavior of RPA module 13416. there is. Adjusting the behavior of the RPA module 13416 may include, for example, changing RPA-based user interface elements presented to users of the smart container system 13000, allowing the RPA module 13416 to interact with one or more external systems. This may include using different robotic process automation technologies to perform features of the RPA module 13416, such as coordinating how to interface, and any other suitable coordination by the RPA module 13416.

In some embodiments, intelligence service 13004 may use data validation information and/or feedback, process validation logs, or a combination thereof as training data. Intelligence service 13004 may be used to influence, adjust, and/or otherwise control the behavior of RPA module 13416 based on data validation information and/or feedback, process validation logs, or a combination thereof. One or more machine learning models can be trained.

In some embodiments, RPA module 13416 may be configured to perform image processing to recognize images in graphical user interfaces that RPA module 13416 interfaces with. RPA module 13416 interfaces may be changed and/or updated using graphical user interfaces of external systems, thereby potentially disrupting robotic process automation-based interfacing with the GUI. The RPA module 13416 can automatically detect changes to the GUI through image recognition and/or image processing. RPA module 13416 may automatically update the robotic process automation-based interfacing with the updated GUI to facilitate continued interfacing with the updated GUI and avoid errors or interruptions in communication with external systems.

In some embodiments, intelligence services 13004 may use image process optimization to use one or more machine learning models to automatically correct robotic process automation-based interfacing with external systems of RPA module 13416. For example, intelligence service 13004 can automatically detect changes in GUIs of external systems and enable RPA module 13416 to automatically continue interfacing with the GUI in light of changes to the GUI. A plurality of GUIs can be used with images as training data to create a machine learning model that can determine how to adjust the robot's process automation.

In some embodiments, RPA module 13416 may be configured to develop a human training system to instruct a human to interface with one or more user interfaces of smart container system 13000 and/or one or more external systems. The human training system may include a plurality of actions and/or techniques used by the RPA module 13416 to interface with one or more user interfaces such that human users can perform tasks similar to the RPA module 13416. Can be taught to one or more human users. A human training system may include one or more documents, videos, tutorials, etc. to facilitate human learning of actions and/or techniques for interfacing with user interfaces.

In some embodiments, RPA module 13416 may be configured to process and document success criteria for robotic process automation implemented by RPA module 13416. Success criteria addressed and documented may be descriptive, so as to facilitate human users of the smart container system 13000 and/or RPA module 13416 interfacing with external systems and/or the smart container system 13000. ) may use processed and documented success criteria to understand one or more process steps and/or algorithms used by the RPA module 13416 to automate the internal market workflows.

In some embodiments, RPA module 13416 may implement gamification of the robotic process automation capabilities of smart container system 13000. Gamification of robotic process automation capabilities includes awarding points to users for performing tasks that are desirable for the operation of the smart container system 13000 and/or for improving the robotic process automation operations of the smart container system 13000. can do. For example, points may be awarded for augmenting robotic process automation algorithms. Users awarded points may compete with each other, and digital and/or physical prizes may be awarded to users who achieve one or more point thresholds and/or rank above one or more other users on a points leaderboard. there is.

In embodiments, smart container system 13000 includes an edge device configured to perform edge computation and intelligence. In some embodiments, edge computing and intelligence may include performing one or both of data processing and data storage in an area physically close to where the processed and/or stored data is needed. In some embodiments, smart container system 13000 may include multiple edge devices. By way of example, an edge device may be a router, routing switch, integrated access device, multiplexer, local area network (LAN) and/or wide area network (WAN) access device, Internet of Things device, smart container, and/or any other suitable device. there is. In some embodiments, edge computation and intelligence may include performing data processing and/or data filtering. The processed and/or filtered data may be transmitted directly to devices that will use the processed and/or filtered data. The processed and/or filtered data may be transmitted along transmission paths that are less congested than general purpose or high traffic data transmission paths. Transmission of processed and/or filtered data may use lower bandwidth than transmission of unprocessed and/or unfiltered data.

In some embodiments, an edge device may implement local edge intelligence to predict relevant delivery factors using data received and/or stored by the edge device. The edge device may be directed to collecting and processing data related to one or more of a particular smart container, a class of smart containers, a shipper, a class of shippers, a shipping line, a class of shipping lines, a container port, a class of container ports, etc. . In some embodiments, the edge device may be physically located near a remote container port or shipping hub area. For example, an edge device may be positioned and configured to collect data regarding performance associated with a specific type of smart container in a geographic area. The edge device may perform data processing, analysis, filtering, trend finding, prediction performance, etc. related to data, and may store processing results, analysis, filtered data, trends, predictions, etc., or parts thereof within the smart container system (13000). Can be dispatched to more centralized servers, processors and/or data centers.

In some embodiments, edge devices may be configured to perform decision making while physically and/or electronically isolated from some or all other components of smart container system 13000. As used herein, electronic isolation may mean or include the temporary inability to communicate with one or more other systems, devices, components, etc. The edge device may make decisions based on outputs and/or conclusions derived from data processing, analysis, filtering, finding trends, performing predictions, etc. related to data received by the edge device. Examples of decisions made by the edge device include whether to verify one or more data, whether to verify a user of smart container system 13000 or a portion thereof, whether a cargo storage and/or transportation service order has been executed, etc. The edge device may transmit data related to decisions made by the edge device to other components of smart container system 13000.

In some embodiments, when an edge device is temporarily electronically isolated from other components of the smart container system 13000, the edge device may make decisions on behalf of other components of the smart container system 13000, and the smart container Smart containers may have decisions audited, evaluated, and/or recorded by other components of the smart container system 13000 when reconnected with other components of the system 13000. Edge devices may be restricted from making some decisions without connection to and/or supervision by other components of the smart container system 13000. Examples of restricted decisions may include decisions related to shipping transactions where confidentiality and/or security are concerned, where volatile cargo is shipped, etc.

In some embodiments, an edge device may store a copy of a distributed ledger, which includes information related to one or more smart containers, smart container fleets, and/or shipping transactions managed by smart container system 13000. The distributed ledger may be a cryptographic ledger such as a blockchain. The edge device can write blocks to a distributed ledger containing smart container information and verify the blocks by comparing them to copies of the distributed ledger stored in other components of the smart container system 13000.

In some embodiments, smart container system 13000 may include a ledger management system configured to manage a network of devices, such as edge devices, that store a copy of the distributed ledger. Devices that store copies of the distributed ledger may be configured to transmit the stored copies to a ledger management system for aggregation, comparison, and/or verification. The ledger management system may establish a whitelist of trusted parties and/or devices, a blacklist of untrusted parties and/or devices, or a combination thereof. The ledger management system can grant permission to specific users, devices, etc. Versions of the distributed ledger can be compared to prevent duplicate transactions, such as the sale of multiple copies of a unique product. In embodiments, smart container system 13000 may be configured to include multiple edge nodes that can each store a copy of the distributed ledger and compare the copies to each other with respect to verification of blocks and addition of new blocks by edge devices and/or all of them. Includes devices.

In some embodiments, smart container system 13000 may implement one or more distributed update management algorithms for updating distributed devices, such as edge devices. A distributed update management algorithm may include one or more procedures for how and when to roll out updates to distributed devices. The smart container system 13000 can manage the version of edge calculation software through a distributed update management algorithm. Distributed devices may receive updates directly from smart container system 13000, may send updates to each other, or a combination thereof.

In some embodiments, smart container system 13000 includes a plurality of edge devices, which may communicate with each other to record and/or verify delivery data. Edge devices may also communicate data related to one or more smart containers, smart container fleets, container ships, container terminals, shipping yards, charging stations, regions, users, shippers, shipping lines, third parties, etc. . An edge device of the plurality of edge devices may communicate such information, possibly in cases where the edge device is electronically isolated from other edge devices and/or other components of the smart container system 13000.

In some embodiments, a first edge device that is electrically isolated and assigned to facilitate smart container repair may be supported by a second edge device. The second edge device determines that the first edge device is not serviceable and/or is not in communication with other components of the smart container system 13000 for an extended period of time so that the possibility of repair by the first edge device cannot be verified. In some cases, the same repair may be assigned. Upon re-entering the communication range, the first edge device may update the second edge device and/or other components of the smart container system 13000 with maintenance operations that occurred while the first edge device was electronically isolated.

In some embodiments, smart container system 13000 includes a hardware failure algorithm configured to determine when one or more components of smart container system 13000, such as an edge device, have stopped operating and/or are otherwise unable to operate properly. It can be implemented. A hardware failure algorithm may include, for example, assigning an edge device to surpass operations previously assigned to a currently malfunctioning or non-operating edge device.

In some embodiments, smart container system 13000 may implement data routing algorithms configured to optimize the flow of data to and/or from edge devices, other components of a fleet system, external systems, or combinations thereof. . The edge device may include one or more signal amplifiers, signal repeaters, digital filters, analog filters, digital-to-analog converters, analog-to-digital converters, and/or antennas configured to optimize the flow of data. In some embodiments, a network enhancement system may include a wireless repeater system such as that disclosed by U.S. Pat. No. 7,623,826 to Pergal, which is incorporated herein by reference in its entirety. The edge device may, for example, filter the data, repeat the data transmission, amplify the data transmission, adjust one or more sampling rates and/or transmission rates, and implement one or more data communication protocols to facilitate the flow of data. can be optimized. In embodiments, an edge device may transmit a first portion of data over a first of a plurality of data paths and a second portion of data over a second of the plurality of data paths. The edge device may determine that one or more data paths, such as a first data path, a second data path, and/or other data paths, are favorable for transmission of one or more portions of data. The edge device may be responsible for one or more types of data being transmitted, one or more protocols suitable for transmission, current and/or expected network congestion, timing of data transmission, current and/or expected volumes of data being transmitted or to be transmitted, etc. Favorable data paths may be determined based on one or more networking variables, such as: Protocols suitable for transmission may include Transmission Control Protocol (TCP), User Datagram Protocol (UDP), etc. In some embodiments, an edge device may be configured to implement a method for data communication, such as that disclosed by U.S. Pat. No. 9,979,664 to Ho et al., which is incorporated herein by reference in its entirety.

In embodiments, smart container system 13000 includes a digital twin module 13420 configured to receive data from an edge device and create a digital twin from the received data. The digital twin generated by the digital twin module 13420 may be one or more digital twins of smart container fleet, smart container, fleet manager, container terminal, loading dock, container ship, shipper, cargo, etc., and may be a digital twin received from the edge device. It can be generated using any or all of the data. Edge devices may transmit vessel-related data, such as data related to smart containers, smart container cargo, shippers, container ports, etc., or a combination thereof. In embodiments, when the smart container system 13000 includes a plurality of edge devices, the digital twin module 13420 may generate a digital twin based on data received from the plurality of edge devices.

In some embodiments, an edge device may be configured to facilitate pre-computation and aggregation of data for a set of user-organized reports. User-configured reports may be integrated into the digital twin created by digital twin module 13420. A user of smart container system 13000 may define one or more parameters of a user-configured report to be included in the digital twin. The edge device may implement one or more data processing and/or filtering depending on the parameters of the user-configured report. The edge device may transmit processed and/or filtered data related to user-configured report parameters to digital twin module 13420. Upon receiving the processed and/or filtered data, digital twin module 13420 may use the received data to create a digital twin that includes a user-configured report and present the digital twin to the user.

In some embodiments, an edge device may be configured to collect and process data for use by one or more artificial intelligence (AI) systems. The AI system may include an intelligence service 13004, one or more artificial intelligence systems configured to facilitate the creation of a digital twin by the digital twin module 13420, and/or any other connected to and/or included in the smart container system 13000. May include artificial intelligence systems. Edge devices may be configured to collect and process and/or filter data to make the data suitable for use by one or more AI systems. An example of processed and/or filtered data collected by an edge device for use by one or more AI systems is training data for use in training one or more machine learning models.

In some embodiments, an edge device may be configured to locally store data related to the creation of a digital twin by digital twin module 13420. In cases where the digital twin is associated with a specific area, shipper, smart container, fleet, container port, vessel, etc., the edge device is, for example, adjacent to the specific area, shipper, smart container, fleet, container port, vessel, etc. By being positioned, it can be specifically positioned to collect and store data for use in populating the digital twin. The edge device may receive, process, filter, organize, and/or store data prior to transmission of the data to the digital twin module 13420, such that the data is associated with and/or related to the population of the digital twin. Suitable. In some embodiments, an edge device may be configured to coordinate the transmission timing of data used to populate the digital twin. The edge device may implement one or more algorithms configured to measure and/or predict congestion of one or more network paths and/or routes, and perform organization of the timing of transmitted data based on the measurements and/or predictions of congestion. can do. The edge device may in some cases prioritize transmission of some types of data over others, such as according to priorities set by the user or by the digital twin module 13420. For example, an edge device may schedule regular transmissions of low-priority information during evening hours when congestion is low, receive high-priority information, and/or transmit high-priority information as soon as it receives a request for high-priority information. Information can be transmitted virtually immediately. In some embodiments, an edge device may be configured to select a data protocol for transmission of data used to populate the digital twin. The edge device may implement one or more algorithms configured to select one or more optimal network paths and/or routes and select a data transmission protocol based on measurements and/or predictions of congestion.

In some embodiments, an edge device can communicate with and receive data from multiple sensors. Edge devices can be configured to intelligently multiplex alternative sensors among those available in the delivery environment for the digital twin.

digital product network

150 illustrates communication between entities within an example digital product network 14000 in accordance with some embodiments of the present disclosure. In some embodiments, digital product networks are communicatively coupled to entities in a value chain network that can provide data related to product level behavior, product level usage, environmental data, data processed at the product level, etc. For example, product data may include sensors, vibration, humidity, temperature, pressure, proximity, level, accelerometer, gyroscope, infrared sensor, MEM, liquid lens, shock, security, mechanical, product, pneumatic, conductivity, and state-dependent frequency monitor. , may relate to at least one of ultrasound, capacitance, or microwaves.

The intelligence layer of the digital product network can then use product-level data to enable companies to address challenges associated with customer demands for quality, efficiency, responsiveness, agility, and transparency. For example, product-level behavior and usage data can be analyzed and manipulated by third-party sources such as artificial intelligence (AI) systems, machine learning (ML) systems, digital twin systems, robotic process automation (RPA) systems, etc. can be combined with Data processing techniques involve at least one of obfuscation, prediction, simulation, transformation, automation, reporting, matching, stream processing, event processing and policy, dispatch and orchestration, analytics and algorithms, or machine learning. It can be.

In some embodiments, the analytics may relate to at least one of descriptive analytics, diagnostic analytics, predictive analytics, or prescriptive analytics. Technical analytics describes what is happening using compressed, accurate, and live data for effective visualization. Diagnostic analytics provides the ability to drill down to the root cause and isolate confusing information. Predictive analytics describes business strategy, proactive adaptation, and simulation. Prescriptive analytics can be evidence-based recommendations to recommend actions and strategies based on challenger testing, and can use advanced AI and analytics techniques to make specific recommendations.

In some embodiments, a digital product network is a network of products that have family relationships with each other. For example, products may all have the same brand, may be part of the same metaverse, may generate data according to the same standardized format, or may be specifically designed to interact or communicate with each other. A digital product network collects information from different parts of a product line to generate information and results that are not available from a single product. A family may include physical products or physical products with mixed content. For example, the content may be technical data, user profile information, or other content.

In some embodiments, analysis and manipulation can lead to valuable insights across the development, manufacturing, supply chain, and customer relationship stages of the product lifecycle. For example, in connected products, new connectivity and technologies enable intelligent provisioning, data aggregation and analytics that can be used to create product connectivity, transactions and support platforms. Data generated by connected products can be analyzed to bridge the gap between supply and demand chains.

Two of the types of interactions that can occur in a digital product network are loosely coupled interactions and platform interactions. In loosely coupled interactions, products are not directly connected to each other. Products operate separately and do not inherently trust their interaction with each other. For example, while cars are on the road, their systems do not rely on data from other cars to control vehicle steering, acceleration, or braking. Products of this type of interaction can use shared intent information, such as when a nearby vehicle indicates that it will soon change lanes. Products that receive and interpret intent can take the intent into account when evaluating response patterns. Data from products with loosely coupled interactions can be fused or integrated for use in artificial intelligence systems, machine learning systems, robotic process automation systems, digital twins, etc.

Platform-based interaction involves different products sharing the same operating platform or ecosystem. These shared ecosystems can coordinate different types of response patterns and encourage or require certain response patterns. For example, smoke detectors in a building may be designed to operate independently, but by sharing a common framework, they can trigger alarms or even critical event responses (eg, a set of sprinklers). Smoke detectors also monitor levels (e.g., frequency of sensor polling, frequency of data transmission, data fidelity) during periods of heightened alert to enable capabilities that might otherwise be dormant for reduced battery consumption. can increase. In the smoke detector example, a smoke detector may be communicatively coupled to an autonomous vehicle system to redirect the vehicle away from fire events. In some embodiments, an autonomous fire truck may drive to a fire event in response to a smoke detector alarm. Such self-responding fire trucks can be advantageous in cases where a community does not have a full-time professional fire department, such as in some rural locations where firefighters can respond to a fire location without first searching for a fire truck.

150, an example of a digital product network service 14002 communicates with and executes algorithms associated with multiple digital entities, such as connected products or intelligent products. For example, digital product network service 14002 may be a version of intelligence service 1IT00 adapted to specific functions of digital product network 14000, described below.

In some example embodiments, connected products may include, among other things, data processing, networking, sensing, autonomous operation, intelligent agents, natural language processing, speech recognition, speech recognition, touch interfaces, remote control, self-configuration, self-healing. , can be enabled by a set of capabilities such as process automation, computation, artificial intelligence, analog or digital sensors, cameras, sound processing systems, data storage, data integration, and/or various Internet of Things (IoT) capabilities. there is. Connected products may involve any form of information technology. Connected products may have processors, computer random access memory, and communication modules. A product can be considered a value chain network entity that fits into the value chain network to provide product usage data.

In the examples provided, connected product 14010, ad hoc network 14012, ad hoc network 14014, local network 14016, and local network 14018 are connected to digital product network services 14002, either directly or through network 14019. communicate with

Connected products may be consumer products, industrial products, or other products with digital components that may communicate with or at least partially include digital product network service 14002. Connected products can be various entities within various industries. For example, each of the connected products may be: Clothing, Electronics (General), Computers and Computer Peripherals, Chemicals (Specialty), Machinery, Food Processing, Auto Parts, Steel, Retail (Online), Retail (Distributor), Retail (Specialty Lines), Retail (General), Retail (Grocery and Food), Electronics (Consumer and Office), Agriculture/Agriculture, Food Wholesalers, or Medical Products.

Connected product 14010 communicates directly with the digital product network service. For example, connected product 14010 may include an antenna for sending electromagnetic band communications directly to an entity hosting digital product network service 14002. In some embodiments, digital product network service 14002 is at least partially hosted within connected product 14010. For example, connected product 14010 may include sensors that send sensor data directly to various intelligence layer modules of digital product network service 14002 that are programmed into processors of connected product 14010.

Ad hoc network 14012 includes connected product 14020 and connected product 14022. Connected product 14020 and connected product 14022 communicate directly with each other and collectively with digital product network service 14002.

Ad hoc network 14014 includes connected product 14024 and connected product 14026. Connected products 14024 and connected products 14026 communicate directly with each other and with digital product network services 14002 via network 14019.

Local network 14016 includes a gateway 14030, at least one sensor system 14032, at least one connected product 14034, and additional data sources 14036. In the example provided, gateway 14030 communicates with sensor system 14032, connected product 14034, and additional data source 14036, respectively. Gateway 14030 then communicates directly with digital product network service 14002. In some embodiments, gateway 14030 hosts digital product network service 14002.

Local network 14018 includes a gateway 14040, at least one sensor system 14042, at least one connected product 14044, and at least one additional data source 14046. Gateway 14040 communicates with sensor system 14042, connected product 14044, and additional data source 14046, respectively. Gateway 14040 then communicates with digital product network service 14002 over network 14019. For example, gateway 14040 may be a router or home automation system hub.

Network 14019 may be any network for communicating data over long distances. In the example provided, network 14019 is the Internet accessed through an Internet Service Provider (ISP).

151 illustrates an example of connected product 14110. In the example provided, connected product 14110 includes a network interface 14112, at least one processor 14114, and at least one memory 14116. In some embodiments, connected products 14010, 14020, 14022, 14024, 14026, 14034, and 14044 have configurations similar to those of connected product 14110.

Connected product 14110 includes at least one network interface 14112, at least one processor 14114, and at least one memory 14116. Network interface 14112 includes one or more communication units that communicate with a network (e.g., Internet, private network, etc.).

The processor 14114 includes a central processing unit (CPU), a graphics processing unit (GPU), a logic board, a chip (e.g., a graphics chip, a video processing chip, a data compression chip, etc.), a chipset, a controller, and a system-on-chip. (e.g., RF system-on-chip, AI system-on-chip, video processing system-on-chip, etc.), integrated circuit, application specific integrated circuit (ASIC), field programmable gate array (FPGA), approximate computing processor, quantum computing processor, It may be any type of computing or processing device capable of executing program instructions, code, binary instructions, etc., including a parallel computing processor, neural network processor, or other type of processor.

In the example provided, processor 14114 includes a data collection module 14120, a data reporting module 14122, a data analysis module 14124, and an intelligence service 14126. For example, modules may be dedicated electronic circuits or non-transitory computer code committed to a computer to instruct a processor to perform the algorithms coded and described herein.

The data collection module 14120 instructs the connected product 14110 to receive data from the sensors and/or network interface 14112 and commit the received data to memory 14116. Data reporting module 14122 retrieves data stored in memory 14116 or redirects data from sensor or network interface 14112 and transmits data from network interface 14112 to a receiving entity. Data analysis module 14124 and intelligence services 14126 perform analysis and run artificial intelligence and/or machine learning algorithms on the data.

Memory 14116 may be any type of non-transitory storage medium, such as one or more of CD-ROM, DVD, memory, hard disk, flash drive, RAM, ROM, cache, network-attached storage, server-based storage, etc. there is. Memory 14116 stores methods, programs, codes, program instructions or other types of instructions that can be executed by processor 14114.

The memory 14116 stores data in various data structures. In the example provided, the data structures include a usage data structure 14130, a sensor data structure 14132, and a derived data structure 14134. Usage data structure 14130 stores data related to the use of connected product 14110. For example, usage data structures can store time, location, settings, and other details about when and how the connected product was used by the user. Sensor data structure 14132 stores data collected from sensors associated with connected product 14110. For example, sensor data may be generated from sensors within connected product 14110 or stored as quantized digital data corresponding to analog sensor signals received at network interface 14112. The derived data structure 14134 stores data derived from usage data and sensor data. For example, data analysis module 14124 may compare frequency of use from usage data with environmental temperature from sensor data to determine frequency of use categorized by environmental temperature range to be stored in derived data structure 14134. You can.

152 illustrates digital product network 14200. In the example provided, a collection of digital products 14210 shares product data 14212 and augmented product data 14214 with intelligence layer 14220. For example, products (including goods and services) may generate data, such as product level data, and transmit it to a communication layer and/or edge data processing facility within the value chain network technology stack. This data can generate augmented product-level data and can be combined with third-party data in an intelligence layer for further processing, modeling or other adaptive or coordinated intelligence activities. This may include, but is not limited to, producing and/or simulating product and value chain use cases, for which data may be used by the product, product development process, product design, etc.

Digital products 14210 may include industrial products 14216, consumer products 14218, and other types of connected products that can communicate with intelligence layer 14220. In some embodiments, digital products 14210 include at least one connected product 14110 of FIG. 151 .

Third-party data system 14221 shares third-party data 14222 with intelligence layer 14220. For example, third party data system 14221 may share third party data 14222 related to any support data used by intelligence layer 14220 that is not yet available in product data 14212. there is.

Intelligence layer 14220 includes various modules that perform analysis on product data 14212, augmented product data 14214, and third-party data 14220. The analysis outputs data 14226 configured to be sent to at least one user system 14228.

In the example provided, intelligence layer 14220 includes demand aggregation module 14230, supply chain management module 14232, new product development module 14234, customer relationship management module 14236, and product lifecycle management (PLM) module 14238. ), digital twin module (14240), synchronized planning module (14242), intelligent procurement module (14244), and dynamic fulfillment module (14246). It should be understood that other embodiments may include other combinations and types of modules without departing from the scope of the present disclosure.

User systems 14228 may be third-party systems that purchase structured data 14226, may purchase analytics results produced by user system 14228, or may be the same enterprise that operates intelligence layer 14220. These may be systems owned by .

In the example provided, user system 14228 includes demand aggregation system 14250, supply chain management system 14252, new product development system 14254, customer relationship management system 14256, and product lifecycle management (PLM) system 14258. ), digital twin system (14260), synchronized planning system (14262), intelligent procurement system (14264), and dynamic fulfillment system (14266). Each of the user systems corresponds to a different use for the configured data, generated by the corresponding module in the intelligence layer.

PLM systems (14238) can provide accurate, up-to-date product information that is accessible throughout the value chain and product life cycle. PLM systems provide augmented cross-function and cross-organizational engagement for design, collaborative innovation, design for manufacturing/procurement, platform-based design philosophy, faster time to market, and improved portfolio management. involvement) can be made possible.

New product development system 14234 is associated with developing and managing product and service value chains that respond to customer experience and are transformed by smart real-time data, advanced technologies, and agile innovation. New product development systems can contribute to improved design quality, increased productivity, and enhanced communication and visibility. For example, simulation that benefits from detailed virtual model processing without expending resources on physically testing the design in a real-world environment; Up-to-date/real-time monitoring of user habits for future design; ease of assembly design; Design for ease of manufacture/collection of parts that will form the product after assembly; Design and compliance with specifications that provide the basic basis for managing operations to produce quality products; and End-to-end transparency, real-time root cause analysis, and proactive resolution driven by customer connectivity for faster problem resolution, problem prevention, customer satisfaction, performance, compliance verification, and avoidance assurance. It can be realized. Additional benefits include supply chain management through the application of the Internet of Things, use of advanced robotics technology, and application of advanced analytics of big data in performing supply chain management and customer satisfaction. In addition to driving other user systems, many additional benefits can influence design decisions. Demand forecasting can be improved with predictive analytics that understand and predict customer demand to optimize supply decisions by enterprise supply chain and business management. Predictive procurement can predict future price trends, price movements, future risks to be managed and required potential through appropriate analysis based on previous procurement data. Real-time/up-to-date product management can improve customer engagement and extend product life. Companies (manufacturers or third-party contractors) can develop the ability to provide services and solutions that complement their existing product offerings (e.g., equipment maintenance, data migration, data storage).

In the customer relationship management system 14236, the integrated solution integrates customer profiles, interactions, and transaction information from multiple applications to provide a view of customers with a solution equipped with industry-specific functionality as well as business intelligence capabilities. can be combined. For example, a network of digital products may include: remote diagnostics that allow subjects to be separated by physical distance instead of subjects being co-located with the person or system performing the diagnosis; inter-machine connectivity (M2) by allowing a sensor or meter to pass the information it records to application software that can use it; and IoT-based digital warranty management systems can enable warranty/repair with proactive and proactive warranty management made easier. Digital Product Networks provide: Brand/Product Agility with real-time product monitoring and the ability to quickly overview the market leading to greater brand agility; Digital product quality with end-to-end transparency, real-time root cause analysis, and proactive solutions driven by customer connectivity; Simulation by providing customers tools to process/use virtual models in VR and AR without expending resources to physically test designs in a real-world environment; Partial ownership use and tracking and physical compliance to meet the parameters of the contract; A common platform and product architecture for a set of stable components that supports evolving diversity and capabilities in the system by limiting connections between other components; and design for consumption insights and methods to improve how customers use/consume products.

Figure 153 illustrates an example digital product network 14300, and Figure 154 illustrates a method 14400 of using product level data according to some embodiments of the present disclosure. Digital product network 14300 is similar to digital product network 14000, where like numbers refer to similar components.

Digital products 14310 are similar to digital products 14210. In the example provided, the digital products include a set of digital products each having a product processor, a product memory, and a product network interface. In some embodiments, one of the digital products 14310 is a product network control tower having a control tower processor, control tower memory, and a control tower network interface. In some embodiments, the product network control tower is a product or server that is not one of digital products 14310. The product processor and control tower processor collectively include non-transitory instructions that program the digital product network system as described herein. For example, intelligence layer 14320 may be distributed between digital product 14310 and remote servers.

In some embodiments, digital products 14310 include a display 14311. In some embodiments, display 14311 is associated with a product network control tower. Display 14311 presents images to the user of display 14311. For example, the display 14311 may be a screen of a mobile phone, television, projector, etc.

In the example provided, a set of digital products and a product network control tower have a set of microservices and a microservices architecture. Digital products 14210 or digital products 14310 generate product level data in a product processor within task 14410 of method 14400. Digital products 14210 or digital products 14310 transmit product level data from the product network interface in task 14412.

The product network control tower receives product level data from the control tower network interface in task 14414. In the example provided, the product processor and the control tower processor are further programmed to communicate based on a shared communication system configured to facilitate communication of product level data from the set of digital products and with the product control tower. In some embodiments, the shared communication system includes electronic devices licensed to an entity that operates or owns digital product network 14300. In some embodiments, the communication system includes a shared security protocol for communicating over shared electromagnetic bands, local area networks, the Internet, 5G, etc.

Digital product network 14300 encodes the product level data into a product level data structure configured to convey the parameters indicated by the product level data across the set of digital products within task 14416. The digital product network 14300 records the product level data structure within the task 14418 in at least one of product memory and control memory. Digital product network 14300 processes product level data structures within task 14420. Digital product network 14300 transmits the processed data to the user system in task 14422.

In some embodiments, intelligence layer 14320 includes a graphical user interface (GUI) module 14340 and a proximity module 14342. The GUI module may generate at least one user interface display for display on display 14311. GUI module 14340 may generate parameters of at least one digitally enabled product of the set of digital products in the at least one user interface display, and a proximal digital product of the set of digital products in the at least one user interface display. You can create close-up displays of products. In some embodiments, generating a proximity display includes generating a proximity display of geographically proximate proximal products, wherein the digital product network selects proximal products by at least one of product type, product capability, or product brand. It is further programmed to filter out the In some embodiments, generating a proximity display includes generating a proximity display of proximity products that are proximate to one of the set of digital products by product type proximity, product capability proximity, or product brand proximity. .

In some embodiments, intelligence layer 14320 includes data integration module 14344. In some embodiments, intelligence layer 14320 includes an edge computation and edge intelligence module or edge module 14346 for edge distributed decision-making between a set of digital products. In some embodiments, edge module 14346 is configured for edge network bandwidth management between or across sets of digital products.

In some embodiments, intelligence layer 14320 includes distributed ledger system module 14348. In some embodiments, the distributed ledger system may be distributed exclusively within digital products 14310. In some embodiments, the distributed ledger is a blockchain ledger.

In some embodiments, intelligence layer 14320 includes a quality management system with a product complaint module 14350 to capture product complaints in a set of digital products. In some embodiments, digital products 14310 detect complaints about other digital products 14310. For example, a digital product may use machine vision or sound processing to identify user frustration while the user is using another digital product.

In some embodiments, intelligence layer 14320 includes product status module 14352. Product status module 14352 may identify the status of a set of digital products. The product status module 14352 may further encode the status as one of the parameters of the product level data structure. Product status module 14352 may also track and/or monitor status across a set of digital products. For example, a bicycle manufacturer could monitor the status of smart bicycles sold to determine potential demand for repair parts or new bicycles. In another example, a rental scooter company may monitor the status of active scooters within its scooter fleet to budget for repairs and replacement scooters.

In some embodiments, intelligence layer 14320 includes a smart contract module 14354 to enable creation of smart contracts based on product level data structures. In some embodiments, intelligence layer 14320 bases the smart contract terms and provisions on a co-location-sensitive configuration of conditions such that they depend on the proximity of a plurality of digital products in the set of digital products. Construct smart contracts.

In some embodiments, intelligence layer 14320 includes robotic program automation (RPA) module 14356. In some embodiments, RPA module 14356 is configured to gamify the interaction based on which digital products are within the set of digital products. In some embodiments, RPA module 14356 generates RPA processes based on the use of a plurality of digital products of the set of digital products.

155 illustrates an example of a digital product network system 14508 where a data enhancement system 14510 receives data from a digital product 14512 for use by a data user system 14514. In some embodiments, data augmentation includes at least one of data fusion and data integration to utilize cross-product data. Data augmentation system 14510 may be part of a control tower, part of a digital twin, or an intelligence service or layer within distributed processors within products, or other suitable systems.

Digital products 14512 are similar to connected products 14110 described above. Each of the digital products 14512 provides data to the data enhancement system 14510. In the example provided, three different digital products 14512 are sharing data with data enhancement system 14510. The data may be usage data for the product, sensor data collected by the product, data retrieved from other products, aggregated data from external sources, or other data obtained through other methods. For example, usage data may include timestamps indicating when the product has been used, length of time data indicating how long the product has been used, data indicating what other products the product has interacted with during use, and any other appropriate usage data. Can contain data. Sensor data may include environmental data, status data, image data, sound data, etc.

Data user system 14514 may be any system that uses or generates augmented data. In some embodiments, data user systems include data enhancement system 14510. For example, a machine learning system can train based on various data streams from products 14512 to generate augmented data. In the example provided, data user systems 14514 include AI/ML systems, robotic process automation (RPA) systems, and digital twin systems.

In some embodiments, digital product network system 14508 is a configuration, home improvement, quality control, or similar system. For example, products 14512 may be part of a family of tools. The laser level can provide leveling data input to handheld self-leveling drills to ensure accurate drill positioning. The digital product network may retrieve relevant specifications (e.g., for appropriate load bearing) provided from the operations control facility. A tool belt can represent appropriate tools to use or access for the next task based on specifications. A tool box with digital capabilities can then indicate the appropriate drill bit and fastener type based on the material of the substrate to be drilled. The self-leveling drill can then retrieve the specifications, set the rotation speed, turn the hammer drill function on or off, etc. Data generated by digital product tools can then be combined to verify proper execution according to workmanship specifications.

In some embodiments, digital product network system 14508 may be an air quality system, an energy audit system, or a similar system. For example, digital products 14512 may be cleaning products, air heating and cooling products, air filter products, window condition detection products, etc. Digital vacuum cleaners can detect the amount of dust and debris picked up during vacuum operation. Data augmentation system 14510 can then fuse the dust data with data from digital heating/cooling and air filter product systems and weather data from third parties. Data user system 14514 can then train AI systems, ML systems, RPA systems, etc. with the fused data to predict indoor air quality metrics and potential causes of poor indoor air quality.

In some embodiments, digital product network system 14508 is a sleep quality system. For example, digital product 14512 may include a digital bed product, a light switch product, a refrigerator door status detector, etc. Digital bed products can detect sleep time, restlessness, and other sleep data. Light switch products can display light status to indicate the amount of ambient light near the bed. The refrigerator door status detector may indicate the last time the door was opened before the user began sleeping in bed. Data augmentation system 14510 may then fuse the sleep data with refrigerator usage data and ambient light data. Data user system 14514 can then train AI systems, ML systems, RPA systems, etc. with the fused data to identify potential foods or eating behaviors that may contribute to poor sleep.

In some embodiments, digital product network system 14508 is an electrical circuit analysis system. For example, digital products 14512 may include circuit breakers digital products, sensitive electronic products, etc. Breaker digital products can generate circuit usage data. Sensitive electronics can generate performance data or input voltage data. Data augmentation system 14510 may then fuse circuit usage data with performance data and voltage input data from various home digital products. The data user system 14514 can then train AI systems, ML systems, RPA systems, etc. with the fused data to map circuits within the building, voltage drops in heavily loaded circuits, etc. Different receptacles to use for the sensitive electronics may be recommended to avoid performance problems from.

In some embodiments, digital product network system 14508 is a children's entertainment management system. For example, digital product 14512 may include a digital children's toy, a television with viewing category data, etc. Digital children's toys can generate usage data. Televisions may generate data indicating the time and duration of children's show viewing. Data augmentation system 14510 may then fuse children's toy usage data with children's viewing shows. Data user system 14514 then uses the fused data to identify the types and characteristics of toys that draw children away from television, market new toys, develop new toys, correct toy focus groups, etc. You can train AI systems, ML systems, RPA systems, etc.

In some embodiments, digital product network system 14508 is a curation system for relating digital entertainment content to augmented reality. For example, digital products 14512 may include televisions, augmented reality headwear, GPS locators, etc. Data enhancement system 14510 may then fuse the television data with location information and landmark information. Data user system 14514 can then train AI systems, ML systems, RPA systems, etc. with the fused data to contextualize augmented reality depictions based on the nature of entertainment choices made on the playback device. . For example, a television show set in city You can. Labels and markers can identify buildings where scenes in a television show occur, such as "This building is where Mafia Boss Z performed his operations." In some embodiments, the system may present material related to botany. For example, augmented reality can curate indicators such as flora-related environmental features. This can be used to curate “tours” of new locations for users. In some embodiments, the system may suggest five routes through the city as the user enters the city. Paths can be supplemented by augmented reality regarding user interests inferred from user media/entertainment consumption. The system can then deliver relevant content to the user's media player, television, book reader, phone, etc.

In some embodiments, digital product network system 14508 is a motion system. For example, digital products 14512 may include a treadmill, exercise bike, stair mill, medication ball with sensor, machine vision product, etc. Exercise products can create records, summaries and analyzes of exercise across a variety of devices. Data augmentation system 14510 may then fuse the data generated across exercise equipment products. Data user system 14514 can train AI systems, ML systems, RPA systems, etc. with fused data to guide users to use the device in a way that complements what the user did with other devices. The system can be matched with a device that understands which muscle groups, calories, etc. are involved or used. For example, if a user rides 10 miles on a bike but has not used a treadmill, weights, step counter, etc., the system could indicate, based on usage information, that the user should perform some exercise on that particular piece of equipment during a given period of time. there is. In some embodiments, the system may be used to monitor patients in cardiac rehabilitation. In some embodiments, the system may monitor athletes for sport-specific improvements, including based on training sets of data by elite athletes across platforms.

In some embodiments, digital product network system 14508 is a carbon footprint calculation system. For example, digital products 14512 may include personal devices or devices that detect actions that contribute to carbon emissions, such as cars, thermostats, household appliances, food purchases (POS data), clothing, etc. Data augmentation system 14510 may then fuse the data generated across carbon footprint calculation products. Data user system 14514 can then train AI systems, ML systems, RPA systems, etc. with the fused data to create plans that meet personal goals, family goals, business goals, regulatory requirements, etc. In some embodiments, the plan implements device controls to limit the use of high carbon emission appliances when their footprint exceeds a threshold.

In some embodiments, digital product network system 14508 is a cross-platform or cross-product reputation system. For example, digital products 14512 may include a variety of digital products that can interact with online communities. Data augmentation system 14510 may then fuse the data generated across digital products. The data user system 14514 can then train AI systems, ML systems, RPA systems, etc. with the fused data to authenticate upright digital citizens and identify digital bad actors. The system may identify cheating, poor behavior, poor sportsmanship, adult language, or other indicators of potential activity that may conflict with terms of use or rules in various online communities, forums, applications, and games. The system can use metaverse IDs, government IDs, credit reports, crime reports, etc. Information can be shared across devices (eg, personal computers, gaming devices, game consoles, gaming handhelds, mobile devices, wearables, virtual reality headsets, etc.) and databases. Metaverse IDs can be associated with government IDs (e.g., state IDs, federal IDs, driver's licenses, passports, etc.) and then also a personal ID across one or more product categories. (e.g., an ID for a website, an ID for an entity, a general metaverse ID) so that bad behavior, cheating, hacking, etc. can be flagged and/or punished across platforms. In some embodiments, the following types of behavior are tracked: posting hateful content/discussions on social media, competitive environments and/or security applications (e.g., cheating programs in video games) , DDoSing websites/apps/game servers, violating terms of service of websites/apps/security databases, system vulnerabilities or money/transaction fraud through exploiting third party programs, etc.). IDs can be used to log in to PCs, laptops, mobile devices, smart watches, smart devices, etc. IDs are tied to network gateways, cellular IDs, or additional other information on the data stream from the device to prevent data from any device, IP address, user, etc. associated with the flagged ID from engaging in certain activities. It can be. Additionally, this traffic and interactions can be throttled, modified, subject to auditing, subject to real-time AI/ML monitoring, etc. Additionally, AI/ML processes can be trained to identify cheats, ToS violations, poor behavior, poor sportsmanship, etc. AI chipsets can be developed and implemented as devices to identify such behaviors, programs, etc. In some embodiments, the system includes incentive programs that provide rewards (e.g., NFTs) for good behavior across products. Rewards and financial details can be implemented as a digital wallet.

In some embodiments, digital product network system 14508 is a personal healthcare system. Digital products 14512 may be a family that includes implantable/permanent medical devices, wearable devices, smartphones, external therapeutic devices, or other health products. Implantable devices provide internal load bearing (such as to measure relative pressures on a joint) as well as data for blood chemistry, blood pressure, tracking immune responses, and other "labs". Data can be generated. Wearable devices can measure and generate data about general health conditions, movement, activity, etc. Smartphone devices can measure and generate data on a variety of user behavioral characteristics, including location and social engagement, affect, happiness, and social metrics. External treatment devices may generate data indicating compliance with medications, physical therapy, and other treatment regimens. The data augmentation system 14510 can then fuse the data generated across the digital products, and the data user system 14514 can then train AI systems, ML systems, RPA systems, etc. with the fused data to enhance the patient's digital It can formulate therapies for twins and simulations, diagnosis, treatment coordination, communication and coaching, proactive problem detection, and more. For example, a personal health care system can provide diagnosis, prescriptions, insurance (underwriting, billing, auditing, payment, reconciliation), treatment (medication, PT, surgery), long-term health care planning and recommendations, and recommendations to improve health (exercise modifications). , social engagement), gamification of health-related behaviors, and other health-related tasks and fields. In some embodiments, the system can monitor sleep patterns, heart rate, blood pressure, and other health parameters combined with data from a smart refrigerator to recommend foods a person should eat to improve overall health.

In some embodiments, digital product network system 14508 is an automotive digital reality system. Digital products 14512 may include automobiles, hearing devices, augmented reality (AR) devices, virtual reality (VR) devices, mixed reality (MR) devices, etc. These families could rotate around a customer's car and seek to leverage data streams from a variety of products that could connect to the vehicle. For example, fusing driving data and in-vehicle passenger observations with audible data including noise, entertainment tracks, and spoken word content can provide a sense of holistic care for the driver or passenger. In-ear devices can be specifically designed to operate as a suite with vehicles. AR/VR headsets will be equipped to help drivers learn to drive a particular car based on the car's current state and configuration as determined by the products, not just about the type or model of the car separate from the current configuration. You can. The AR/VR headset will allow users to play social video games, such as a game that allows car owners to race their cars with friends, based on data indicating current tire condition, fuel level, location, etc. It can be configured to do so.

In some embodiments, the digital product network system 14508 is a family of products with elements configured to reside and operate in digital wallets, metaverses, AR/VR devices, vehicles, individual rooms in the home, workplaces, smart cities, nature, etc. .

In some embodiments, digital product network system 14508 is an industrial system. Digital products 14512 include machine sensors that indicate the need for adjustments to the machine (e.g., the need to increase fan usage for a machine that is too hot, the need for lubrication, materials in terms of manufacturing or packaging systems). the need to repair broken or missing rollers indicated by a conveyor belt sensor, etc.). Data augmentation system 14510 can fuse sensor data within the warehouse for specific purposes, such as environment (e.g., temperature, airflow, humidity, lighting, UV light, etc.). This fusion may relate to each of these types of data, or a combination of such data may be fused on separate devices such that the results from the analytics engine can only be output to the mobile device based on thresholds. (e.g. too cold - suggests heat; too much humidity - start a dehumidifier or start a humidifier when too dry; pull down shades if too much UV light; close windows for increased airflow) open, turn on fans, etc.). This environmental data can be fused with sensor data about the manufacturing system or systems for packaging devices. This data can be individually fused to an analysis engine that provides results of machine states. Data user system 14514 then compiles the results of the combination of machine sensing and environmental sensing into a software application on the mobile device to provide recommendations to address environmental issues or machine condition issues and/or automate systems. AI systems, ML systems, RPA systems, etc. can be trained with fused data to create output analysis engines, recommendation engines, and/or automation engines. The system can also monitor a combination of environmental conditions and machine health conditions to determine optimal conditions for maximum output to automatically adjust machine and environmental conditions to provide optimal conditions while keeping costs to a minimum. .

In some embodiments, digital product network system 14508 is a sports equipment system. Digital products 14512 may include golf clubs and golf balls, baseball bats and baseballs, hockey sticks and hockey pucks, tennis racquets and tennis balls, bowling balls, etc. Various hitting tools and balls can generate data that indicates how far, how straight they travel, and how the ball travels to the target. The machine vision system or sensors of the striking tool may generate data about acceleration, angle, rotation, and other data about the swing of the striking tool. The data augmentation system 14510 can then fuse the data generated across the sports equipment, and the data user system 14514 can use the fused data to train AI systems, ML systems, RPA systems, etc. to coach and swing the user. Corrections to improve can be identified. In some embodiments, the system fuses data from smart exercise machines and smart watches/wearables.

In some embodiments, digital product network system 14508 is a physical retail system. Digital products 14512 may include products to be purchased, product packaging, shopping carts, smartphones, etc. The product may represent the type of item added to the cart, the quantity of the item added to the cart, the item on the smartphone's shopping list, etc. The data augmentation system 14510 can fuse data, and the data user system 14514 trains an AI system, ML system, RPA system, etc. with the fused data, performs suggestions for complementary products, and stores You can display the locations of complementary products, provide incentives, etc. For example, incentives may include discounts, reward points, digital badges, etc.

In some embodiments, digital product network system 14508 is a commercial lending risk management system. Digital products 14512 may include products in warehouses, packaging, environmental sensors, or similar products. Products may generate data indicating the proximity of different products in the same warehouse or the presence of the same or different products in different warehouses. The data augmentation system 14510 can fuse data, and the data user system 14514 trains AI systems, ML systems, RPA systems, etc. with the fused data to inform borrowers of risks or disadvantages regarding loan (credit limit) conditions. You can indicate compliance, charge prices into different brackets, or alert of movement of inventory items that may trigger additional charges. The system can also flag extreme hazards that may be of interest to government authorities, such as hazardous or explosive materials stored or moved under hazardous conditions.

In some embodiments, digital product network system 14508 is a media consumption recommendation system. Digital products 14512 may include microphones, televisions, etc. The product can generate data that indicates the music a person listens to, a person's taste in books, or a person's interest in television and movie content. The data augmentation system 14510 can fuse data, and the data user system 14514 can train AI systems, ML systems, RPA systems, etc. with the fused data to recommend video games that appeal to people. there is.

In some embodiments, digital product network system 14508 is a health improvement system. Digital products 14512 include networked exercise equipment, rowing machines, stationary bikes, smart devices (e.g., refrigerators), wearables (e.g., smart rings or smart watches), smart beds, and other May include products. The products generate data that can be used to better understand the user's overall health. The data augmentation system 14510 can fuse data, and the data user system 14514 trains an AI system, ML system, and RPA system with the fused data to determine the user's eating habits, exercise habits, and sleeping/sitting habits. You can learn about The system can determine whether the user is consuming too many calories or eating the wrong foods considering their exercise routines. The system may also track the user's sleep patterns and determine whether the user is exercising and/or eating at appropriate times. Networked exercise equipment may be owned by a user or may be owned by a gym. When the equipment is placed in the gym, exercise devices can pair with the user's phone or wearable device to know who is on the equipment, how long they have been using it, etc. Smart refrigerators may include sensors and imaging devices that determine what a user purchases, what he or she actually eats, and when he or she eats it. In some embodiments, the smart refrigerator may include a voice control interface that verifies the user's profile (e.g., family with children, single, living together but no children, etc.) and/or who is eating, what is cooking, etc. Includes. The data can train AI systems to determine a user's overall health profile. The system can be configured to make recommendations to the user, such as better foods, better meal times, how much to exercise, how many times to exercise, etc.

In some embodiments, digital product network system 14508 is a product maintenance system. Digital products 14512 may include any product or machine with some level of connectivity and some components of warranties/repair. The resulting data sets and cumulative data layers can be analyzed and used for remote diagnosis, warranty pricing, repair pricing, replacement suggestions, upgrade suggestions, etc. In the case of remote diagnosis, the subject may be separated by physical distance instead of being in the same location as the person or system performing the diagnosis. In the case of warranty and repair pricing, proactive and proactive warranty management becomes easier through IoT-based digital warranty management system.

In some embodiments, digital product network system 14508 is a power consumption management system. Digital products 14512 may include thermostats, light switches, light bulbs, refrigerators, coffee makers, HVAC products, etc. A power consumption monitor divided by breakers in an electrical service panel box can be connected to a variety of appliances/circuits. In some embodiments, the products provide enough data that a circuit panel box monitor can be bypassed to view product-specific usage patterns. A single circuit can be monitored to understand different appliance consumption patterns (refrigerator on/off, vacuum on/off, HVAC on/off/heat/cooling/fan) and other smart devices (thermostats, light switches/bulbs, etc.) ) may be the source-protection (parent-umbrella) for. The system can combine data and perform analysis of usage patterns (refrigerator temperature settings, coffee maker on times, thermostat/HVAC usage, etc.) for maintenance, behavioral usage guidance, and more. For example, HVAC can monitor airflow by power consumption (taking into account temperature, humidity, etc.) to recommend filter changes or behavior suggestions to save energy.

In some embodiments, digital product network system 14508 is a work management system. Digital products 14512 may include location sensors, status sensors, task completion sensors, etc. The data augmentation system 14510 can fuse data, and the data user system 14514 can fuse data into an AI system, ML system, or RPA system with the fused data to coordinate tasks and/or services in the metaverse, real life, or both. You can train your back. For example, the system may use geographic location, couponing, logistics, etc. to determine which employees, family members, or other participants are best positioned to perform a task. The best positions can be pre-programmed or AI-optimized for priorities such as savings, timing, importance, and more. The system may provide task sequencing, task command, and coordination of other participants accordingly. The system can operate based on tasks to be performed in real life and in metaverse work environments, and can link metaverse and real life activities. In some embodiments, the system may put services up for bidding. In some embodiments, the system provides a marketplace for services within the metaverse and a metaverse working environment for metaverse workers and real workers to work cooperatively to achieve results.

In some embodiments, digital product network system 14508 is a personal health system. Digital products 14512 include wearable devices, eyeglasses, glasses with liquid lenses, active clothing (heating, cooling, stress/strain, force, etc.), recreational devices (GPS monitor devices, etc.), buildings and other environmental systems. (temperature, humidity, lighting, etc.), automotive systems, etc. The product generates sensors or other defined data (micro or macro) that affects the health of the individual. Data augmentation system 14510 can fuse multiple data streams, including those outside of ecosystems, that can be integrated to monitor and manage personal health. The data user system 14514 can train AI systems, ML systems, RPA systems, etc. with fused data to augment and verify specific analysis models (GPS navigation algorithms, etc.). The system may provide activity alerts, automated reporting that builds a personal profile over time, suggested activity measurements, instructions for performing activity measurements (e.g., by activating clothing or other devices, etc.) there is. The system includes psychological assessment, recommendations, patient referral, identification of risk factors such as temperature, UV exposure, presence of pathogens, etc., cumulative exposure to carcinogens or other things that may cause long-term disease, lifetime medical analysis, and temperature changes. Clothing activation, massage, etc. can be performed.

Futures Smart Contract

156 illustrates a smart futures contract system 15000. Smart futures contract system 15000 may, for example, connect various value chain network entities (e.g., products with intelligent features, packaging or containers with intelligent features, infrastructure or facilities with intelligent features, A set of smart contracts associated with (transportation systems, planning systems, etc.) with intelligent features, such as managing or mitigating risks (e.g. shortages of supply, supply chain disruptions, changes in demand, prices of inputs) provide operating efficiencies (e.g., by hedging or providing improved results in the event of a variety of potentially adverse contingencies such as changes in market prices, changes in market prices, etc.), and provide operational efficiencies (e.g., based on plans or forecasts). (e.g., by ensuring the availability of items), improving revenues (e.g., by obtaining inputs at more favorable prices than would otherwise be available, etc.), and/or, e.g. A set of smart contracts configured to provide different benefits by participating in the futures markets associated with the set (including various markets for options, futures, etc. involving commodities, stocks, currencies, energies, and other items). may be related to Items accompanying smart contracts include the goods, services, and any mixed product containing components of various types of goods and services described herein and in documents incorporated herein by reference. can do.

In some embodiments, robotic process automation may operate with demand-side planning to orchestrate futures contracts. For example, a robotic agent obtains commitments for supply upon finding market conditions that satisfy a set of parameters or conditions (e.g., pricing conditions) set in de-risking algorithms. for a set of inputs that will be required to provide a planned inventory set of items (e.g., parts, components, fuel, materials, or many others), so that a set of smart futures contracts can be automatically executed. A set of de-risking algorithms may be performed to construct terms and conditions for a set of smart futures contracts that set prices, delivery times, and delivery locations. The robot agent is equipped with a set of expert procurement experts, including contracts recommended or engaged by these experts under these conditions, and demand planning inputs (e.g. demand forecasts, inventory forecasts, etc., demand elasticity curves, competitive behavior). can be trained based on a training set of data, such as a training set of interactions with a set of inputs (such as predictions, supply chain forecasts, and many others). This may include the interaction of these experts within the enterprise demand planning software suite. Agents may use demand factors, supply factors, price factors, and other factors, including those that generate expected balances between supply and demand, appropriate inventory estimates, recommendations for price, location, and timing for supply and/or distribution, etc. It may be trained to interact with or include a set of demand planning models, such as models that predict factors. In embodiments, de-risking algorithms are intended to reduce various risks and contingencies, including those mentioned above, such as shortages of supply, supply chain disruptions, changes in demand, changes in prices of inputs, and changes in market prices. , macroeconomic factors, geopolitical disturbances, disruptions caused by weather and climate, the impact of epidemics or pandemics, counterparty risk (including anti-money laundering risk, credit risk, default risk, etc.), etc. may include In embodiments, de-risking algorithms may include input biases (e.g., biases in training inputs, biases in models, biases due to incomplete or inaccurate data, etc.), biases in weights, and other factors. Algorithms attempt to mitigate the risks created by the use of other algorithms, such as those that help identify various biases that may include those where algorithmic performance is inferior to human performance (e.g., where intelligent systems This may include those that identify where some important elements of the system cannot be effectively replicated. In embodiments, these de-risking algorithms may provide a set of recommendations for smart futures contracts and/or adjustments to the de-risking algorithms used to construct smart futures contracts. In some embodiments, smart futures contract system 15000 operates from, implements, or is integrated with a digital twin (e.g., a supply chain digital twin or a generic digital twin interface).

In some embodiments, smart contract system 15000 constructs and/or enters a set of smart future contracts with futures system 15006 based on conditions in the value chain network. In some embodiments, smart futures contract system 15000 operates on a value chain network based on conditions and prices within a set of futures markets.

In some embodiments, smart futures contract system 15000 may be at least partially integrated into a product or product packaging to manage or mitigate risk. For example, if a product or product packaging is exposed to adverse environmental conditions, a smart futures contract could automatically be configured as an option to acquire a set of replacement products, covering the contingency that the product has incurred damage that would require replacement. You can. This may, for example, determine whether a replacement will actually be required while the product/package is still in transit as determined by sensors on the product, package, transport vehicle, or nearby infrastructure. Testing a set of products well enough is possible or close to happening before this happens. The construction of these option-type futures contracts may be based on models or predictive artificial intelligence systems (e.g. , generated by algorithms that can be trained based on historical data sets and other inputs), as well as predictions of the impact of replacement needs (including the impact of delays and/or supply reductions on prices and other factors). It can be based on In embodiments, smart futures contracts may be structured with appropriate option durations to allow for determination of the actual extent of demand for replacement, appropriate option prices, etc., thereby mitigating the risk of catastrophic losses while mitigating the likelihood of profitable outcomes. is maintained to the extent possible under the circumstances. In embodiments, an option type futures contract to acquire a substitute product may provide alternatives to the substitute, such as providing a refund to the customer, providing a substitute product or service, providing an incentive to accept the delayed product, etc. It can be paired with a set of automatically constructed futures contracts that mitigate risk by setting conditions for. These contracts may be constructed using inputs, models and algorithms similar to those used in other smart futures contracts described herein.

In some embodiments, smart futures contract system 15000 renegotiates the set of futures prices based on current market conditions. Renegotiation may be performed by a set of robotic process automation agents or other artificial intelligence systems, such as those trained, based on historical data, based on feedback from results, and/or based on human interactions involved in contract negotiations. It can be performed by As one of many examples, upon recognition of a possible widespread supply chain disruption for an input component for a set of goods, the system may determine future prices of inputs (e.g., to ensure continuity of supply), future prices of outputs (e.g., to ensure continuity of supply), may propose to renegotiate (e.g., to reflect possible increases in market prices), and other factors, which may be proposed in a set of futures smart contracts that implement the proposed terms of the renegotiation.

In some embodiments, smart futures contract system 15000 may, optionally with the aid of analytics, models, etc. based on historical procurement data, determine future price trends, price movements, future risks to manage, and potentially Involves or undertakes predictive procurement to anticipate different elements required. These include weather, climate, geopolitical situations, epidemics/pandemics, counterparty behavior, government behavior (including import and export regulations and their enforcement), transportation, port congestion, inventory levels of key components, availability and prices of materials and many others. A model that takes these factors into account may be included.

In the example provided, smart futures contract system 15000 includes at least one contract entity 15001, at least one data source 15002, at least one intelligent service 15004, at least one futures system 15006, and at least Includes one distribution system (15008). The contract entity 15001 is an entity in the value chain network that owns, leases, leases, purchases, or otherwise controls the intelligent service 15004. For example, contracting entity 15001 may be a manufacturer of a product interested in managing raw material scarcity risk in future product cycles. In another example, contracting entity 15001 may be a clothing manufacturer interested in early identification and price negotiation for fabrics, dies, designs, etc. that may become popular for the next fashion season. In another example, contracting entity 15001 is an industrial entity that monitors the status of components in various machines and orders for future delivery of components that are predicted to fail in the near future. It can be. In the industrial entity example, the contract entity 15001 is "parting out" a machine rather than repairing it when futures prices indicate that it is less expensive to part out a new machine than to make ongoing repairs. ), the intelligent service 15004 can be configured to compare futures prices for purchasing new machines with futures prices for selling non-failed components of the machines.

Data source 15002 generates data for use by intelligence service 15004. Data may be measured locally (e.g., by sensors or IoT devices), retrieved from third parties, determined or augmented by other intelligence services, or in any other way, without departing from the scope of this disclosure. can be collected. For example, data source 15002 may generate product level data, customer level data, or data at other value chain levels. In the example provided, data sources 15002 include directly connected customers 15010, intelligent products 15012, and environmental sensors 15014. Intelligent products 15012 and environmental sensors 15014 include vibration, humidity, temperature, pressure, proximity, level, accelerometers, gyroscopes, infrared sensors, optical sensors, MEMS, liquid lenses, shock, security, Sensor data may be provided associated with any measurable parameter, such as at least one of a machine, product, pneumatic, conductive, state-dependent frequency monitor, ultrasonic, capacitance, or microwave. Data sources 15002 may include Internet of Things (IoT), social networks, social media, automated agent behavior, business entity behavior, human behavior, data source results, data source parameters, wearables, personal, financial, economic, credit score, environment. , weather, labor, employment, census, crime, health, living, journalism/media, entertainment, location/motion, loyalty, reputation, real estate, reviews, marketing, food and drugs, education, Data may be generated associated with any suitable topic or industry, such as at least one of retail, transportation, biometrics, travel, events, or customer activity.

Intelligence service 15004 may be a configured version of intelligence service 1IT00. For example, intelligence service 15004 may be adapted to execute certain functions of smart futures contract system 15000 described below. In the example provided, intelligence service 15004 includes at least one of data storage 15019, smart contract service 15020, demand aggregation service 15022, digital wallet 15024, risk determination service 15026, and robotic process automation ( RPA) service (15028).

Data storage 15019 may be any type of non-transitory storage medium, such as one or more of CD-ROM, DVD, memory, hard disk, flash drive, RAM, ROM, cache, network-attached storage, server-based storage, etc. It can be. Data storage 15019 stores methods, programs, codes, program instructions, or other types of instructions that can be executed by processors of intelligent service 15004.

The data storage 15019 stores data in various data structures. In the example provided, data storage 15019 includes a risk data structure 15021 and a robotic process automation (RPA) data structure 15022. Risk data structure 15021 stores risk tolerance information, risk identification, risk assessment, and other risk information associated with contract entity 15001. For example, risk data structure 15021 may store price change tolerance and the maximum price that contracting entity 15001 may be interested in paying for a product in the future.

RPA data structure 15022 stores robotic process automation algorithms executed by RPA service 15028. Algorithms may use various artificial intelligence embodiments described throughout this disclosure and documents incorporated herein by reference, such as various types of neural networks, various algorithms and expert systems, etc.

Smart contract service 15020 generates execution of smart contracts based on analysis from data from data sources 15002, information received from futures systems 15006, and other features of intelligence service 15004. , modify, and monitor. For example, the smart contract service 15020 may specify the type of commodity, the number of units (e.g., barrels of oil, bushels of wheat, ounces of gold, etc.), and the commodity in a smart contract governing a futures contract for a commodity. The contract price to be paid, the execution date of the futures contract, and other appropriate parameter values may be specified. In some embodiments, smart contract service 15020 may be configured to: initiate an authentication process associated with a transaction, initiate a reporting process associated with a transaction, configure logistics information associated with a transaction, set conditions (e.g. , premium rate, interest rates, contract price, delivery date, payment due date, etc.) can display parameter values corresponding to triggering actions. It should be understood that the types of data that can be used to parameterize a smart contract may be different without departing from the scope of this disclosure.

In some embodiments, smart contract service 15020 may operate autonomously. For example, smart contract service 15020 may operate without human intervention based on instructions from RPA service 15028 or other criteria provided by contract entity 15001.

In some embodiments, demand aggregation service 15022 may predict that demand for certain raw materials used in the manufacture of products may increase in the near future based on data from environmental sensors 15014. . Based on the predicted increase in demand, the smart contract service 15020 can create a smart contract with the futures system 15006 for future delivery of raw materials in anticipation of a price increase due to the predicted increase in demand, and data Other data from source 15002 suggests that production of raw materials may not keep pace with expected increases in demand.

Smart contract service 15020 receives evidence of completion of a task to trigger actions (e.g., payment, recording, etc.) in response to completed tasks. As another example, smart contract service 15020 may monitor futures prices and buy or sell goods and services based on the risk tolerance indicated by the contract entity.

The demand aggregation service (15022) aggregates demand across groups, locales, etc. Aggregating may include aggregating demand for at least one of virtual products, virtual events, virtual services, or services associated with virtual products or events. Demand aggregation service 15022 monitors demand response across multiple systems, such as how demand responds to changes in supply (e.g., scarcity effects), price changes, customization, pricing, advertising, etc.

In some embodiments, demand aggregation service 15022 may provide information, orders, and/or commitments for one or more products, categories, raw materials, components, logistics reservations, supplies, equipment, etc., optionally implemented as one or more contracts, which may be smart contracts. ) is counted. Demand aggregates may include current demand for existing products and future demand for products that are not yet available.

Digital wallet 15024 stores banking and financial information for payments into and out of futures system 15006. For example, digital wallet 15024 may store cryptocurrency information, bank balance and routing information, and other information that can be used to complete transactions in futures and distribution systems 15008.

Risk determination service 15026 determines the risk associated with events and conditions received by data source 15002, demand aggregation service 15022, and futures system 15006. For example, Risk Decision Service 15026 may determine that the financial risk of waiting to purchase raw materials when needed is greater than the risk of purchasing raw materials from the futures system 15006 at some point before the raw materials are needed for delivery. Data can be provided to the smart contract service (15020). In some embodiments, risk determination service 15026 retrieves risk information, such as price fluctuation tolerance, from risk data 15021. If the price fluctuation tolerance limit is low, the smart contract service 15020 can execute a smart contract for the forward of a commodity as the futures price approaches the maximum price that the contract entity 15001 is willing to pay. Due to the high price fluctuation tolerance limit, the smart contract service 15020 may allow the futures price to exceed the maximum price in the hope that the price may fall later.

Robotic process automation (RPA) service 15028 may facilitate, among other things, computer automation of creating and validating smart contracts between contract entity 15001 and futures system 15006. In some embodiments, RPA service 15028 monitors human interaction with various systems to learn the patterns and processes performed by humans in the performance of individual tasks. This may include observation of human behaviors involving interactions with hardware components, software interfaces, and other elements. Observations are made in real-world testing by humans, such as when humans tag or label training data sets with features that help RPA systems learn to recognize or classify features or objects, among many other examples. may include field observations of a human performing an action, as well as observations of other activities or simulations of a human performing an action with the explicit intent of providing a training data set or input for the RPA system.

In some embodiments, the RPA service 15028 can learn to perform certain tasks based on learned patterns and processes, such that tasks can be performed in support of or on behalf of a human decision maker. 15028). For example, RPA service 15028 may identify that a farmer typically schedules truck services to transport crops to a sales location approximately two weeks before harvesting the crops. The RPA service 15028 may further identify that the crop has grown to a certain height approximately three weeks before the farmer performs annual services on the harvesting machine and harvests the crop. On the basis of detecting the harvester's annual service (e.g. by data from the machine itself, by identifying fuel filters on credit card receipts, by machine vision identifying service, etc.) or (e.g. by data from Based on identifying that the height of the crop has reached a certain height (by machine vision data from source 15002), RPA service 15028 may query logistics reservation system 15038. Based on the response from the logistics reservation system 15038, the smart contract 15020 may present reservation options to the farmer for trucking services at three weeks from service or crop height determination, or negotiate a reservation on the farmer's behalf. .

A future system 15006 may be any system in which goods and services to be delivered or performed in the future are purchased and sold. For example, the futures system 15006 may involve forward contracts, exchange futures, options, various derivatives, etc. Goods and services to be delivered or performed may include real estate, goods, raw materials, finished goods, accounting services, or any other material or performable service that may create an obligation to be delivered or performed in the future. In the example provided, futures system 15006 includes component futures (15030), commodity futures (15032), consumer goods futures (15034), equipment futures (15036), and logistics reservations (15038).

Component gift 15034 may relate to mechanical parts, repair parts of a product, wearable parts of a product, upgrade parts of a product, parts to be assembled into a new product by the manufacturer, and other component types. In some embodiments, component gifts 15030 are associated with circular economy systems. For example, the smart contract service 15020 may perform circular economy optimization based on the futures price of a commodity such as a component.

Commodity futures (15032) can be materials used to create other commodities. For example, the raw material 15032 may be copper, steel, iron, lithium, etc. Consumer goods futures 15034 may relate to items consumed when creating a product or performing a service. For example, consumer products may include razor blades, coffee pods, disposable batteries, raw pork belly, etc.

Equipment gifts (15036) may relate to machinery, vehicles, and other equipment. Logistics reservation gift (15038) may relate to gift services for warehousing, transportation, etc. For example, logistics reservation gifts 15038 may include port docking reservations, truck reservations, warehouse space rentals, canal passage reservations, etc.

Distribution system 15008 relates to at least one of picking, packing, moving, storing, receiving, transporting, or delivering a set of items in a supply chain. In the example provided, smart contract service 15020 coordinates with the delivery dates and locations detailed in smart contracts for delivery, storage, and handling of physical items through distribution system 15008. and contracts for other handling of items are entered into the logistics reservation system 15038.

In some embodiments, the value chain may include an intelligent agent system receiving feedback from users regarding each intelligent agent. For example, in some embodiments, a client application utilizing an intelligent agent may provide an interface through which a user can provide feedback regarding actions output by the intelligent agent. In embodiments, the user provides feedback identifying and characterizing any errors by the intelligent agent. In some of these embodiments, a report may be generated (e.g., by a client application or platform) indicating the set of errors encountered by the user. Reports can be used to reconfigure/retrain intelligent agents. In embodiments, reconfiguring/retraining an intelligent agent includes removing inputs that are a source of error, reconfiguring a set of nodes of an artificial intelligence system, reconfiguring a set of weights of an artificial intelligence system, and reconfiguring the set of outputs, reconfiguring the processing flow within an artificial intelligence system (e.g., placing gates on a recurrent neural network to balance learning with the need to reduce certain inputs to avoid explosive error problems) RNNs), reengineering types of artificial intelligence systems (e.g., convolutional neural networks, recurrent neural networks, feedforward neural networks, long/short-term memory (LSTM) neural networks, self-organizing neural networks, or many other types. and, in combination, by modifying the type of neural network), and/or augmenting the set of inputs to the artificial intelligence system.

In an embodiment, a library of neural network resources representing combinations of neural network types that mimic or simulate neocortex activity is provided to human experts to perform a variety of activities that are the subject of the development of intelligent agents, such as those involving robotic process automation. It can be configured to allow the selection and implementation of modules that replicate the combinations used by. In embodiments, the various neural network types from the library may be configured in serial and/or parallel configurations to represent the processing flow, when involved in an activity that is the subject of automation, e.g., based on spatiotemporal imaging of the brain. It can be arranged to mimic or replicate the flow of processing in . In embodiments, an intelligent software agent for agent development may select a set of neural network resource types, arrange the neural network resource types according to a processing flow, and configure the neural network resource types, such as using any of the training techniques described herein. Configure input data sources for a set of neural network resources, and/or automatically deploy a set of neural network types on available computational resources to initiate training of a set of neural network resources configured to perform the desired intelligent agent/automation workflow. can be trained to do so. In embodiments, an intelligent software agent used in agent development operates on an input data set of spatiotemporal imaging data of the human brain, such as an expert performing a workflow that is subject to the development of additional applications, and initiates learning of a neural network. It uses spatiotemporal imaging data to automatically select and organize the selection and arrangement of sets of types. Accordingly, a system for developing intelligent agents may be configured for (optionally automatic) selection of neural network types and/or arrangements based on the spatiotemporal neocortical activity patterns of human users involved in the workflow in which the agent is trained. Once developed, the resulting intelligent agent/process automation system can be trained as described throughout this disclosure.

In embodiments, a system for developing intelligent agents (including the aforementioned agents for development of intelligent agents) may be configured to enable a human user to (optionally automatically) infer which data sources should be selected as input to the intelligent agent. Information from brain imaging is available. For example, for processes in which the neocortical area (O1) is highly active (involving visual processing), visual input (e.g., available information from cameras, or especially visual representations of information such as price patterns) is advantageous data. Can be selected as a source. Similarly, for processes involving area C3 (which involves storage and retrieval of facts), a data source that provides reliable factual information (such as a blockchain-based distributed ledger) may be selected. Accordingly, systems for developing intelligent agents can be configured for (optionally automatic) selection of input data types and sources based on the spatiotemporal neocortical activity patterns of human users involved in the workflow for which the agent is trained.

157 illustrates an example environment of an edge networking (EDNW) system 16100 in accordance with some embodiments of the present disclosure. In embodiments, edge networking system 16100 provides a framework for providing edge networking services to one or more edge environments 16102-108. In some embodiments, edge networking service 16100 may be at least partially replicated in each of edge environments 16102-108. Examples of edge environments 16102-108 in which edge networking system 16100 can be at least partially replicated include devices 16102, premises 16104, telecom installations 16106, and computing clouds 16108. In these embodiments, an individual instance of edge networking system 16100 may include some or all of the capabilities and/or modules of edge networking system 16100 discussed herein, thereby creating an edge networking system ( 16100 is adapted to the specific functions performed by each edge environment 16102-108 to which edge networking system 16100 is replicated. Examples of edge devices 16102 include client devices, smart devices, and connected devices, wearable devices, and any other suitable device. Examples of premises 16104 include home/residential buildings, factory buildings, office buildings, campuses, government buildings, medical buildings, and any other suitable building, campus, or premises. Examples of edge telecom installations 16106 include base stations, satellites, regional data centers, relay stations, signal arrays, and any other suitable telecom installation. Examples of computing clouds 16108 include in-cloud compute networks, private clouds, public clouds, hybrid clouds, multiclouds, and infrastructure-as-a-service cloud. ) clouds, platforms-as-a-service clouds, software-as-a-service clouds, and any other suitable type of computing clouds.

Additionally or alternatively, in some embodiments, edge networking system 16100 may be implemented as a set of microservices such that different edge environments 16102-108 are exposed to edge environments 16102-108. Edge networking systems can be leveraged through one or more APIs. In these embodiments, edge networking system 16100 can be configured to perform various types of edge networking services that can be adapted to different edge environments 16102-108. In either of these configurations, the edge environment 16102-108 can provide a networking request to the edge networking system 16100, whereby the request can perform a specific networking task (e.g., initiate a connection, initiate a series of connections, Encryption of data, transmission of encrypted data over the connection, use of protocols for connection and/or data transmission, routing decisions or calculations, use of application-specific protocols for data transmission, AI chipset interfacing instances, 5G software-defined, AI assisted or enabled networking, tunable signal filtering, AI-assisted or enabled network augmentation, or digital twin network configuration, simulation, prediction, testing, etc.). In response, edge networking system 16100 executes the requested networking task for each edge environment 16102-108.

Additionally or alternatively, in some embodiments, edge networking system 16100 may be implemented using one or more specialized chips comprised of microservices and/or networking tasks. In embodiments, edge networking system 16100 may communicate via VCN bus 16110. One or more of the edge environments 16102-108 may be connected to a VCN bus, thereby allowing instances of edge networking system 16100 to communicate with each other over the bus. VCN control tower 16112 may also be connected to the bus. As such, one or more instances of edge networking system 16100 may transmit data to and receive signals from VCN control tower 16112.

Figure 158 illustrates an example embodiment of an environment of edge networking system 16100 where multiple VCN system services 16202-206 are connected to VCN bus 16110. In embodiments, edge networking system 16100 may be configured to support a VCN control tower and/or one of other VCN system services 16202-206 (e.g., energy systems and processes such as DPNW systems, ROBO systems, NRGY systems, etc.). Data may be transmitted to and/or data received from one or more of them. Energy systems and processes may be referred to as energy systems (e.g., NRGY systems), which may be or include energy systems, processes, modules, services, platforms, etc., described in this disclosure. Data received by edge networking system 16100 from VCN control tower 16112 may be, for example, related to VCN tasks ordered by the VCN control tower and to be completed by sharing of data between multiple VCN entities. Can contain data. Edge networking system 16100 may perform one or more networking tasks in accordance with one or more VCN tasks, such as encryption codes, routing data, protocol information, data, AI predictions, AI routing information, AI conclusions, or any other suitable data. can be performed. Edge networking system 16100 may additionally or alternatively transmit data to other of the VCN entities via the VCN bus.

In some embodiments, edge networking system 16100 is configured to facilitate optimization of processes and communication between different VCN modules. Edge networking system 16100 may be configured to communicate and/or encrypt data packets, for example, between two of the different VCN modules, and/or between one or more data sources and another of the VCN modules. By optimizing the protocols, one or more networking tasks can be performed in conjunction with one or more of the other VCN modules. For example, a VCN task requiring high-speed transmission of data from IoT sensors to one or more modules of the ROBO system, as well as high-speed reporting of energy systems by the NRGY system to the ROBO system, may require VCN control tower 16112 to transmit data from IoT sensors to one or more modules of the ROBO system. This may involve instructing the edge networking system 16100 to optimize data routes and protocols between the system and one or more of the IoT data sources and the NRGY system.

159 illustrates an example embodiment of an edge device 16102 that includes an instance of edge networking system 16100. 159 illustrates the edge networking system 16100 installed on device 16102, the edge networking system and/or configured instances of the edge networking system may be installed on any suitable type of edge environment 16102-108 disclosed herein. It will be appreciated that it can be uploaded/installed/operated/operated on or within. An instance of edge networking system 16100 may include one or more modules configured to facilitate performance of networking tasks by edge networking system 16100. Instances of edge networking system 16100 may be specifically altered, added, removed, and/or otherwise adapted to suit the desired performance of the instance of edge networking system 16100 for the particular edge environment in which the configured instance of edge networking system 16100 is stored. Modules that can be customized otherwise. For example, different configurations of edge networking system 16100 may be implemented based on each of the different types of edge environments 16102-108, as well as in some cases individual edge environments within the types of edge environments 16102-108. can be configured and uploaded/installed based on Configurations of individual instances of the edge networking system may be determined by VCN control tower 16112 according to one or more of the VCN tasks.

In embodiments, edge networking system 16100 may include one or more of the following modules: edge device-as-a-service module 16302, an application-specific protocol module (16304), Edge Robotics Module (16306), SDWAN Module (16308), Network Customization Module (16310), AI Chipset Interfacing Module (16312), 5G Software Defined Module (16314), Edge Networking AI Module (16316), Tunable Signal filtering module (16318), network routing module (16320), AI network enhancement module (16322), and edge network digital twin module (16324).

In embodiments, edge device-as-a-service module 16302 may configure one or more functions of an edge environment 16102-108 in which edge networking system 16100 is installed and/or one or more functions of a configured edge networking device 16100. It is configured to facilitate provisioning of other devices and/or platforms with which the edge networking system 16100 communicates, such as via the VCN bus 16110. For example, a configured instance of the edge networking system 16100 installed in a smart container may utilize the functionality of the smart container to other connected items (e.g., a fulfillment system) through the edge device-as-a-service module 16302. It can be made possible.

In embodiments, application-specific protocol module 16304 may be configured to enable edge environments 16102-108 in which edge networking system 16100 is installed and/or enable edge networking system 16100 to communicate, e.g., via VCN bus 16110. and configured to determine and/or facilitate provisioning of one or more application-specific networking protocols to one or more other devices and/or platforms. For example, application-specific protocol module 16304 may facilitate communication to and/or from applications requiring data to be sent and/or received over proprietary encrypted communication protocols.

In embodiments, edge robotics module 16306 is configured to facilitate networking related to robotics in edge environments 16102-16108. Edge robotics module 16306 may perform edge networking functions, for example, routing one or more robotics-related data, selecting and enabling robotics-related communication protocols, local control interfacing, remote control interfacing, sending and receiving sensor data, and other suitable features. The system 16100 may provide for robots local to and/or remote from the edge environment 16102-108 in which it is installed.

In embodiments, SDWAN module 16308 is configured to facilitate creating, managing, and/or handling communications over a software-defined wide area network (SDWAN). Edge networking system 16100 may define SDWAN, for example, via SDWAN module 16308, and/or some or all of the devices connected to SDWAN may be defined by one or more SDWAN modules 16308. Instances of edge networking system 16100 may be installed thereon to facilitate communications within the SDWAN.

In embodiments, network customization module 16310 is configured to facilitate creation, management, customization, and/or handling of one or more networks. Edge networking system 16100 may customize the network for specific VCN tasks through network customization module 16310. For example, a network customization module may facilitate customization of networks associated with VCN tasks involving predicting demand for specific products and/or industries. The network can be customized to facilitate fast, efficient and secure communication between devices connected to it.

In embodiments, AI chipset interfacing module 16312 is configured to enable performance of one or more networking tasks by edge networking system 16100 via one or more AI chipsets. For example, the edge networking system 16100 may make one or more network routing decisions through an AI system built on an AI chipset. Additionally or alternatively, the AI chipset interfacing module 16312 receives data from an AI chipset that communicates with an edge environment 16102-108 in which the edge networking system 16100 is installed, such as via a VCN bus 16110. /or may be configured to facilitate transmitting data to the AI chipset.

In embodiments, 5G software-defined module 16314 is configured to facilitate management, customization, handling, and/or other tasks related to 5G networks. For example, the telecom installation 16106 in which the edge networking system 16100 is installed may be or include a 5G base station, and the software defined module 16314 of the edge networking system 16100 of the 5G base station may be configured to provide May perform one or more networking tasks related to managing traffic. Additionally or alternatively, in some examples, 5G-enabled mobile devices connected to a 5G network may have communication over the network facilitated by instances of 5G software defined module 16314 installed thereon.

In embodiments, edge networking AI module 16316 is configured to perform artificial intelligence and/or machine learning related functions related to one or more of the networking tasks. For example, routing decisions, protocol selections, protocol management, encryption processes, filtering decisions, and/or any other suitable type of networking task may be performed by AI/ML via edge networking AI module 16316. and/or may be assisted.

In embodiments, tunable signal filtering module 16318 is configured to perform tuning of digital signals transmitted and/or received by and/or within the edge environment 16102-108. Tunable signal filtering module 16318 may include one or more tunable digital filters. Tunable digital filters tune signals to reduce network traffic, for example, by culling unnecessary communications, and eliminate “piggybacking” of third-party signals on the edge environment 16102-108. Or you can reduce, filter IoT sensor data, etc.

In embodiments, network routing module 16320 is configured to perform network routing operations on signals transmitted and/or received by edge environment 16102-108. Network routing operations include, for example, determining the optimal data path for sensor data within the premises, determining optimal traffic flow for a 5G network, determining the routing device to which mobile data should be transmitted, etc. May contain the same networking tasks.

In embodiments, AI network augmentation module 16322 is configured to optimize network performance through one or more AI and/or machine learning processes. For example, AI network augmentation module 16322 may utilize one or more AI and/or machine learning methods to determine optimal protocols for data throughput, and/or perform predictions and/or simulations of network conditions and congestion/throughput. You can use the process.

In embodiments, the edge network digital twin module 16324 is configured to generate and/or generate one or more digital twins associated with networking by, in, and on an edge environment by one or more of the edge environments 16102-108. It is configured to manage. For example, digital twin module 16324 can model the network within a manufacturing plant and run simulations via the digital twin to predict network congestion.

In some embodiments, edge network system 16100 has a system for decoupling congestion control from link loss.

In some embodiments, edge network system 16100 has an intelligence layer on top of UDP.

In some embodiments, edge network system 16100 has an automated policy and governance engine for edge workload placement.

In some embodiments, edge network system 16100 has edge data integration with a service-oriented architecture.

In some embodiments, edge network system 16100 has edge-specific protocols.

In some embodiments, edge network system 16100 has an edge storage protocol.

In some embodiments, edge network system 16100 has an edge storage protocol integrated with AI managed storage.

In some embodiments, edge network system 16100 has a distributed edge database.

In some embodiments, edge network system 16100 has an edge-distributed query language.

In some embodiments, edge network system 16100 has an edge policy engine.

In some embodiments, edge network system 16100 has an automated edge data marketplace.

In some embodiments, edge network system 16100 has an RF filtering system for wireless nodes.

In some embodiments, edge network system 16100 has network coding.

In some embodiments, edge network system 16100 has a machine learning and/or artificial intelligence system at each node of the network.

In some embodiments, edge network system 16100 has a machine learning and/or artificial intelligence system at each node of the network configured for filtering and multi-level signal compression based on signal characteristics.

In some embodiments, edge network system 16100 has a machine learning and/or artificial intelligence system at each node of the network configured to filter based on context and/or content.

In some embodiments, edge network system 16100 has a machine learning and/or artificial intelligence system at each node of the network configured to filter based on feedback on the results.

In some embodiments, edge network system 16100 has a machine learning and/or artificial intelligence system configured to archive at an optimal level of granularity.

In some embodiments, edge network system 16100 has an AI-augmented edge-aware network fabric.

In some embodiments, edge network system 16100 has a machine learning and/or artificial intelligence system configured for optimization of storage capacity.

In some embodiments, edge network system 16100 has a machine learning and/or artificial intelligence system configured for optimization of compute.

In some embodiments, edge network system 16100 has a machine learning and/or artificial intelligence system configured for optimization of energy.

In some embodiments, edge network system 16100 has a mail protocol for mail.

In some embodiments, edge network system 16100 has a video streaming protocol for video streaming.

In some embodiments, edge network system 16100 has a gaming protocol for gaming.

In some embodiments, edge network system 16100 has a system for monitoring packet activity.

In some embodiments, edge network system 16100 has a system for storing and replicating data streams for later use by machine learning and/or artificial intelligence systems.

In some embodiments, edge network system 16100 has robotics as edge devices.

In some embodiments, edge network system 16100 has a system to enable local control.

In some embodiments, edge network system 16100 has a system for customizing the security of the network and/or devices on the network.

In some embodiments, edge network system 16100 has a system for configuring channels for various types of communications.

In some embodiments, edge network system 16100 has a system for configuring the environment after discovery of devices.

In some embodiments, edge network system 16100 has a system for customizing the network by data scheduling and resource availability.

In some embodiments, edge network system 16100 has a system for customizing data routing, processing, and/or computing based on data type and/or security.

In some embodiments, edge network system 16100 has custom virtualization deployments.

In some embodiments, edge network system 16100 has a system for customization of neural network layering to form new AI structures.

In some embodiments, edge network system 16100 has a system for customizing antenna layouts and power delivery to meet specified needs.

In some embodiments, edge network system 16100 has a system for enabling wireless charging including AI-based battery optimization and AI management of heat generation.

In some embodiments, edge network system 16100 has a system for customizing the human interface.

In some embodiments, edge network system 16100 has a system for customization of optimal routing algorithms in service-tailored 5G networks.

In some embodiments, edge network system 16100 has a network device.

In some embodiments, edge network system 16100 has an AI chipset.

In some embodiments, edge network system 16100 has channelization beyond the front end.

In some embodiments, edge network system 16100 has channelization beyond the interface.

In some embodiments, edge network system 16100 has a system for creating and managing digital twins of network and/or AI edge systems.

In some embodiments, edge network system 16100 has an edge device-as-a-service system.

In some embodiments, edge network system 16100 has a reader-as-a-service system.

In some embodiments, edge network system 16100 has a gateway-as-a-service system.

In some embodiments, edge network system 16100 has a repeater-as-a-service system.

In some embodiments, edge network system 16100 has an asset tag-as-a-service system.

In some embodiments, edge network system 16100 has a robotic data collector-as-a-service system.

Quantum Computing for VCNs

160 illustrates an example quantum computing system 17000 according to some embodiments of the present disclosure. In embodiments, quantum computing system 17000 provides a framework for providing a set of quantum computing services to one or more quantum computing clients. In some embodiments, the quantum computing system 17000 framework may be at least partially replicated at each quantum computing client (e.g., VCN control towers and/or various VCN entities). In these embodiments, an individual client may include some or all of the capabilities of quantum computing system 17000, whereby quantum computing system 17000 can perform certain functions performed by subsystems of the quantum computing client. Adapted to the field. Additionally or alternatively, in some embodiments, quantum computing system 17000 may be implemented as a set of microservices such that different quantum computing clients can access the quantum computing system (17000) via one or more APIs exposed to the quantum computing clients. 17000) can be used. In these embodiments, quantum computing system 17000 may be configured to perform various types of quantum computing services that can be adapted to different quantum computing clients. In either of these configurations, a quantum computing client may provide a request to quantum computing system 17000, whereby the request is to perform a particular task (e.g., optimization). In response, the quantum computing system 17000 executes the requested task and returns a response to the quantum computing client.

160, in some embodiments, quantum computing system 17000 includes a quantum adaptation service library 17002, a quantum general service library 17004, a quantum data service library 17006, and a quantum computing engine library 17008. , a quantum computing configuration service 17010, a quantum computing execution system 17012, and a quantum computing API interface 17014.

In embodiments, quantum computing engine library 17008 includes quantum computing engine configurations 17016 and quantum computing process modules 17018 based on various supported quantum models. In embodiments, quantum computing system 17000 may include a quantum circuit model, a quantum Turing machine, a spintronic computing system (e.g., non-e.g., such as those using diamond materials). using spin-orbit coupling to generate spin-polarized electronic states in magnetic solids, adiabatic quantum computers, unidirectional quantum computers, quantum annealing, and various quantum cellular automata. Many different quantum models can be supported, including but not limited to (quantum cellular automata). Under the quantum circuit model, quantum circuits can be based on quantum bits, or “qubits,” which are somewhat similar to the bits of classical computation. Qubits can be in a 1 or 0 quantum state, or in a superposition of 1 and 0 states. However, when the qubits measure the result of a measurement, they will always be in the 1 or 0 quantum state. The probabilities associated with these two outcomes depend on the quantum state the qubit was in immediately before the measurement. Computations are performed by manipulating qubits with quantum logic gates, which are somewhat similar to classical logic gates.

In embodiments, quantum computing system 17000 may be physically implemented using analog or digital methods. Analog approaches may include, but are not limited to, quantum simulations, quantum annealing, and adiabatic quantum computation. In embodiments, digital quantum computers use quantum logic gates for computations. Both analog and digital approaches can use quantum bits, or qubits.

In embodiments, quantum computing system 17000 includes a quantum annealing module 17020 that uses quantum perturbations to achieve a given goal over a given set of candidate solutions (e.g., candidate states). It can be configured to find the global minimum or maximum of an objective function. As used herein, quantum annealing refers to the use of quantum fluctuation-based computations instead of classical computations to obtain measurements such as size, length, cost, time, distance or other measurements from within a possibly very large but finite set of possible solutions. May refer to a meta-procedure for finding a procedure that identifies an absolute minimum or maximum value. The quantum annealing module 17020 can be used to solve problems with many local minima and discontinuities in the search space, such as finding the ground state of a spin glass or the traveling salesman problem (e.g., a combinatorial optimization problem). ) can be used.

In embodiments, quantum annealing module 17020 starts from a quantum-mechanical superposition of all possible states (candidate states) with equal weights. Quantum annealing module 17020 can then evolve and follow the natural quantum-mechanical evolution of systems (e.g., physical systems, logical systems, etc.), e.g., the time-dependent Schrödinger equation. In embodiments, the amplitudes of all candidate states are varied to realize quantum parallelism depending on the time-dependent strength of the transverse field, resulting in quantum tunneling between states. If the rate of change of the transverse field is sufficiently slow, the quantum annealing module 17020 can remain close to the ground state of the instantaneous Hamiltonian. If the rate of change of the transverse field is accelerated, the quantum annealing module 17020 may temporarily leave the ground state, but may create a higher probability of terminating in the final matter energy state or the ground state of the Hamiltonian.

In embodiments, quantum computing system 17000 may include any number of qubits and transport ions to spatially distinct locations within an array of ion traps, forming a remotely entangled ion chain. Large entangled states can be constructed through photonically connected networks.

In some implementations, quantum computing system 17000 includes a trapped ion computer module 17022, which can be a quantum computer that applies trapped ions to solve complex problems. The trapped ion computer module 17022 can have low quantum decoding and can construct large solution states. Ions, or charged subatomic particles, can be trapped and suspended in free space using electromagnetic fields. Qubits are stored in the stable electronic states of each ion, and quantum information can be transferred through the overall quantized motion of the ions in shared traps (which interact through the Coulomb force). The laser can be applied to induce coupling between qubit states (for single-qubit operations) or between internal qubit states and external motion states (for entanglement between qubits). there is.

In some embodiments, a traditional computer, including a processor, memory, and a graphical user interface (GUI), can be used to design, compile, and provide output from execution, and quantum computing system 17000 can execute machine language instructions. Can be used to run In some embodiments, quantum computing system 17000 can be simulated by a computer program executed by a traditional computer. In these embodiments, a superposition of states of quantum computing system 17000 can be prepared based on input from initial conditions. Because the initialization operations available in quantum computers can only initialize a qubit to either the |0> or |1> state, initializing it with a superposition of states is physically unrealistic. However, for simulation purposes, it may be useful to bypass the initialization process and initialize quantum computing service 17000 directly.

In some embodiments, quantum computing system 17000 provides various quantum data services including quantum input filtering, quantum output filtering, quantum application filtering, and quantum database engine.

In embodiments, quantum computing system 17000 may include quantum input filtering service 17024. In embodiments, quantum input filtering service 17024 may be configured to select whether to run a model on a quantum computing system 17000 or a classical computing system. In some embodiments, quantum input filtering service 17024 may filter data for later modeling on a classical computer. In embodiments, quantum computing system 17000 may provide input to traditional computing platforms while filtering out unnecessary information from flowing to distributed systems. In some embodiments, system 17000 can trust through filtered specific experiences to intelligent agents.

In embodiments, a system within one of the systems may deploy quantum computing or quantum algorithmic resources to a value chain network activity, traditional computational resources and algorithms, or a hybrid thereof, based on a set of inputs. It may include a model or system to automatically decide whether to apply the combination. In embodiments, inputs to the model or automated system may include demand information, supply information, energy cost information, capital costs for computational resources, development costs (e.g., for algorithms), energy costs, and operating costs. (including labor and other costs), performance information on available resources (quantum and traditional), and (e.g., a wide variety of simulations described herein and/or in documents incorporated herein by reference). may include any of a number of different data sets that can be used to simulate (using any of the techniques) and/or predict the difference in results between quantum-optimized and non-quantum-optimized results. there is. A machine learning model (included in the DPANN system) uses input data for a given request to determine which set of resources to deploy in a given way, either by deep learning the results or from human expert decisions. Can be trained on a data set of The model may itself be deployed on quantum computational resources and/or use quantum systems, conventional systems, and/or hybrids or combinations, such as quantum annealing, to determine whether, where, and when. Quantum algorithms can be used.

In some embodiments, quantum computing system 17000 may include quantum output filtering service 17026. In embodiments, quantum output filtering service 17026 may be configured to select a solution from solutions of multiple neural networks. For example, multiple neural networks can be configured to generate solutions for a particular problem, and quantum output filtering service 17026 can select the best solution from the set of solutions.

In some embodiments, quantum computing system 17000 connects and directs the neural network development or selection process. In this embodiment, quantum computing system 17000 can directly program the weights of the neural network so that the neural network provides desired outputs. This quantum-programmed neural network can then operate without the supervision of quantum computing system 17000, but will still operate within the expected parameters of the desired computational engine.

In embodiments, quantum computing system 17000 includes a quantum database engine 17028. In embodiments, quantum database engine 17028 is configured for in-database quantum algorithm execution. In embodiments, a quantum query language may be used to query the quantum database engine 17028. In some embodiments, the quantum database engine prioritizes quantum workflows, including prioritization of query workloads, such as based on overall priority as well as comparative advantage of using quantum computing resources versus others. There may be an embedded policy engine 17030 for allocation and/or allocation. In embodiments, quantum database engine 17028 may assist in the recognition of entities across value chain networks by establishing a single identity for what is valid across interactions and touchpoints. Quantum database engine 17028 may be configured to perform optimization of data matching and intelligent traditional compute optimization to match individual data elements. The quantum computing system 17000 may include a quantum data obfuscation system for obfuscating data.

Quantum computing system 17000 may include, but is not limited to, analog quantum computers, digital computers, and/or error-corrected quantum computers. Analog quantum computers can directly manipulate interactions between qubits without decomposing these actions into primitive gate operations. In embodiments, quantum computers capable of running analog machines include, but are not limited to, quantum annealers, adiabatic quantum computers, and direct quantum simulators. Digital computers can operate by performing an algorithm of interest using primitive gate operations on physical qubits. Error-corrected quantum computers can refer to a version of gate-based quantum computers made more robust through the deployment of quantum error correction (QEC), which allows noisy physical qubits to emulate stable logical qubits. ), allowing the computer to behave reliably for arbitrary calculations. Additionally, quantum information products may include, but are not limited to, computing power, quantum predictions, quantum optimizations, and quantum decision support.

In some embodiments, quantum computing system 17000 is configured as an engine that can be used to optimize traditional computers, integrate data from multiple sources into a decision-making process, etc. The data integration process may involve real-time capture and management of interaction data with extensive tracking capabilities that are both directly and indirectly related to value chain network activities. In embodiments, quantum computing system 17000 may store cookies, email addresses and other contact data, social media feeds, news feeds, event and transaction log data (transaction events, network events, computational events, and many others), event streams, results from web crawling, distributed ledger information (including blockchain updates and state information), results from distributed or federated queries of data sources. , streams of data from chat rooms and discussion forums, and many other things.

In embodiments, quantum computing system 17000 includes a quantum register with a plurality of qubits. Additionally, the quantum computing system 17000 may include a quantum control system for implementing basic operations for each of the qubits in the quantum register and a control processor for coordinating the necessary operations.

In embodiments, quantum computing system 17000 is configured to optimize pricing of smart container-based freight transportation services. In embodiments, quantum computing system 17000 may utilize quantum annealing to provide optimized freight transportation service pricing. In embodiments, quantum computing system 17000 may use q-bit based calculation methods to optimize price.

In embodiments, quantum computing system 17000 is configured to optimize design or construction features of value chain network products, devices, vehicles, services, etc. For example, the quantum computing system 17000 may be configured to optimize product design, smart container design, robot design, smart container fleet configuration, robot fleet configuration, liquid lens design, data story configuration, etc. Additionally or alternatively, quantum computing system 17000 is configured to optimize movement or paths of value chain network entities, including robots or robot fleet paths, smart containers or smart container fleet paths, etc.

In embodiments, quantum computing system 17000 is configured to automatically discover smart contract construction opportunities. Automated discovery of smart contract composition opportunities (e.g., by robotic process automation (RPA) of stakeholder, asset, and transaction types) can be based on marketplaces and published APIs for machine learning. there is.

In embodiments, quantum-established or other blockchain-based smart contract applications may include booking a set of robots from a robot fleet, booking a smart container from a smart container fleet, and agreeing a transfer price between subsidiaries. This may include, but is not limited to, executing them. In embodiments, quantum-established or other blockchain-enabled smart contracts enable frequent transactions to occur between a network of parties, and manual or replication operations are performed by counterparties for each transaction. A quantum-established or other blockchain acts as a shared database to provide a single, secure source of truth, and smart contracts automate authorizations, calculations, and other transaction activities that are prone to delays and errors. Smart contracts may use software code to automate tasks, and in some embodiments, this software code may include quantum code to enable highly sophisticated optimized results.

In embodiments, quantum computing system 17000 or another system within one of the systems may include a quantum-enabled or other hazard identification module configured to perform hazard identification and/or mitigation. Steps that can be taken by the risk identification module may include, but are not limited to, risk identification, impact assessment, etc. In some embodiments, the risk identification module determines a risk type from a set of risk types. In embodiments, risks may include, but are not limited to, preventable, strategic, and external risks. Preventable risks can refer to risks that arise internally and can usually be managed at a rules-based level, such as by monitoring operating procedures, using employee and manager guidelines and specifications. Strategic risk can refer to risks voluntarily taken to achieve greater rewards. External risks can refer to risks that occur externally and are not within the company's control (for example, natural disasters). External risks are unpreventable or undesirable. In embodiments, a risk identification module may determine expected costs for any category of risk. The risk identification module can perform calculations of current and potential impacts on the overall risk profile. In embodiments, a risk identification module may determine the probability and severity of a particular event. Additionally or alternatively, the risk identification module may be configured to anticipate events.

In some embodiments, quantum computing system 17000 or another system of system 17000 is configured for accelerated sampling from probabilistic processes for risk analysis. In embodiments, quantum-simulated accelerated testing is initiated to maintain accelerated life tests with constant stress loadings, including accelerated degradation tests and time-varying stress loadings.

In embodiments, quantum computing system 17000 or another system of system 17000 is configured for graph clustering analysis for anomaly and fraud detection.

In some embodiments, quantum computing system 17000 includes a quantum prediction module configured to generate predictions. Additionally, the quantum prediction module can configure classical prediction engines to further generate predictions, reducing the need for ongoing quantum computation costs, which can be significant compared to traditional computers.

In embodiments, quantum computing system 17000 processes input vector data if the covariance matrix of the data is efficiently obtainable as a density matrix, under certain assumptions about the vectors given in quantum mechanical form. It may include a quantum principal component analysis (QPCA) algorithm that can be used. It may be assumed that the user has quantum access to the training vector data in the quantum memory. Additionally, each training vector can be assumed to be stored in quantum memory in terms of its difference from the class means. Such QPCA can be applied to provide dimensionality reduction using the computational advantages of quantum methods.

In embodiments, quantum computing system 17000 is configured for graph clustering analysis for certified randomness on proof-of-stake blockchains. Quantum cryptography techniques can use quantum mechanics in their designs, which allows such techniques to rely on supposedly unbreakable laws of physics for their security. Quantum cryptography techniques can be information-theoretically secure so that their security is not based on arbitrary non-fundamental assumptions. Information-theoretic security is not proven in the design of blockchain systems. Rather, classic blockchain technologies typically rely on security claims that make assumptions about the limitations of an attacker's resources.

In embodiments, quantum computing system 17000 is configured to detect adversarial systems, such as adversarial neural networks, including adversarial convolutional neural networks. For example, quantum computing system 17000 or another system of platform 17000 may be configured to detect fake trading patterns.

In embodiments, quantum computing system 17000 includes a quantum continual learning (QCL) system 17032, where QCL system 17032 continuously and adaptively learns about the external world to: Updating the quantum model to account for different tasks and data distributions enables autonomous, progressive development of complex skills and knowledge. QCL system 17032 operates on realistic time scales where data and/or tasks become available only during operation. Previous quantum states can be superimposed within the quantum engine to provide capacity for QCL. Because the QCL system 17032 is not limited to a finite number of variables that can be treated deterministically, it can continuously adapt to future states, creating dynamic continuous learning capabilities. QCL system 17032 may have applications where data is received continuously, although data distributions remain relatively static. For example, QCL system 17032 can be used in quantum recommendation applications or quantum anomaly detection systems where data is continuously received and the quantum model is continuously refined to provide various results, predictions, etc. QCL enables asynchronous alternate training of tasks, updating the quantum model only on real-time data available from one or more streaming sources at a specific moment.

In embodiments, QCL system 17032 operates in complex environments where target data is constantly changing based on uncontrolled hidden variables. In embodiments, QCL system 17032 can scale in terms of intelligence while processing increasing amounts of data and maintaining a realistic number of quantum states. The QCL system 17032 applies quantum methods to allow execution of successive computations to provide detailed-driven optimal results, while greatly reducing the requirement for storage of historical data. In embodiments, QCL system 17032 is configured for unsupervised streaming perception data because it continuously updates the quantum model with new available data.

In embodiments, QCL system 17032 enables multi-modal-multi-task quantum learning. The QCL system 17032 is not limited to a single stream of perceptual data, but allows many streams of perceptual data from different sensors and input modalities. In embodiments, QCL system 17032 can solve multiple tasks by replicating quantum states and executing computations on the replica quantum environment. A key advantage of QCL is that the quantum model does not need to be retrained on historical data because the superposed state retains information related to all previous inputs. Multimodal and multitask quantum learning enhances quantum optimization because it gives quantum machines reasoning skills through the application of vast amounts of state information.

In embodiments, quantum computing system 17000 supports quantum superposition, or the ability of sets of states to be overlaid on a single quantum environment.

In embodiments, quantum computing system 17000 supports quantum teleportation. For example, information can be transferred between photons in chipsets even if the photons are not physically connected.

In embodiments, quantum computing system 17000 may include a quantum transfer pricing system. Bilateral transfer pricing allows the establishment of prices for goods and/or services exchanged between commonly controlled entities that are subsidiaries, affiliates or part of a larger corporation, and can be used to provide tax savings for the entities. A quantum transfer pricing system is configured to use quantum computing techniques to solve the transfer pricing problem across all systems within one of the systems and interfaces connecting the systems. In embodiments, solving the transmission pricing problem involves testing the elasticity of each system within one of the systems with a set of tests. In these embodiments, testing can be done in periodic batches and then repeated. As described herein, transfer pricing may refer to the price charged by one business unit of a company to another business unit of the company for goods and services. In embodiments, a quantum transfer pricing system may be applied across a value chain network to optimize overall product value.

In embodiments, a quantum transfer pricing system integrates all financial data related to transfer pricing on an ongoing basis for all entities in an organization, where the integration overlays the data into a single quantum state. ) involves applying quantum entanglement to In embodiments, financial data may include profit data, loss data, data from branch office invoices (potentially including quantity and price), etc.

In embodiments, a quantum transfer pricing system may interface with a reporting system that reports segmented P&Ls, transaction matrices, tax optimization results, etc. based on overlapping data. In embodiments, a quantum transfer pricing system automatically generates prediction calculations and evaluates expected local benefits for an arbitrary set of quantum states.

In embodiments, the quantum transfer pricing system may be integrated with a simulation system to perform simulations. The proposed optimal values for new product prices can be discussed cross-border through integrated quantum workflows and quantum transport communication states.

In embodiments, quantum transfer pricing may be used to proactively control the distribution of revenue within a multi-national enterprise (MNE), e.g., over the course of a calendar year, such that the entities may each It is possible to achieve arms-length profit ranges for this type of trade.

In embodiments, the QCL system 17032 may utilize a quantum comparable uncontrolled price (QCUP) method, a quantum cost plus percent method (QCPM), a quantum resale price method (QRPM), a quantum transaction net margin method (QTNM), and quantum profit splitting. Quantum transfer pricing can be calculated using a number of methods, including methods.

The QCUP method can potentially apply quantum computation to find comparable transactions made between related and unrelated organizations through the sharing of quantum overlapping data. Through the application of a bilateral comparison engine, a benchmark price can be determined by comparing the price of goods and/or services in an internal transaction with the price used by independent parties.

The QCPM method can compare gross profit to cost of sales and thus measure cost-plus mark-up (the actual profit earned from the product). Once this markup is determined, it must be identical to what a third party would make for a comparable transaction in a comparable context with similar external market conditions. In embodiments, the quantum engine may simulate external market conditions.

The QRPM method looks at groups of transactions rather than individual transactions and is based on the gross profit margin, or difference, between the price at which the product was purchased and the price at which it was sold to a third party. In embodiments, a quantum engine may be applied to calculate price differences and record transactions in the nested system.

The QTNM method is based on the net proceeds of controlled transactions rather than comparable external market prices. The calculation of net profit is done through a quantum engine that takes into account a wide variety of factors and can optimally resolve the product price. The net profit can then be compared to that of an independent company, potentially using quantum teleportation.

The quantum profit-split method can be used when two related companies work on the same business venture, but separately. In these applications, quantum transfer pricing is based on revenue. The bilateral profit split method applies quantum calculations to determine how the profits associated with a particular transaction are divided among the independent parties involved.

In embodiments, one of the systems includes quantum-aware device-level kits, quantum-aware industrial Internet of Things (IoT) kits, quantum-enabled FPGAs, and different quantum computer types and/or Can support quantum-aware device stacks that include systems with recognition of the capabilities of different quantum algorithm types.

In embodiments, quantum computing system 17000 may utilize one or more artificial networks to fulfill requests from quantum computing clients. For example, quantum computing system 17000 may identify patterns in images (e.g., using image data from a liquid lens system), perform binary matrix factorization, perform local content targeting, and A set of artificial neural networks can be utilized to perform similarity-based clustering, collaborative filtering, opportunity mining, etc.

In embodiments, one of the systems may include a hybrid computing allocation system for prioritizing and allocating quantum computing resources and traditional computing resources. In embodiments, the prioritization and allocation of quantum computing resources and traditional computing resources is measure-based (e.g., measuring the degree of advantage of a quantum resource over other available resources). , it can be cost-based, optimality-based, speed-based, impact-based, etc. In some embodiments, the hybrid computing allocation system is configured to perform time division multiplexing between quantum computing system 17000 and traditional computing system. In embodiments, a hybrid computing allocation system may automatically track and report allocation of computational resources, availability of computational resources, cost of computational resources, etc.

In embodiments, quantum computing system 17000 may be utilized for queue optimization for use of quantum computing resources, including context-based queue optimizations.

In embodiments, quantum computing system 17000 may support quantum-computation-aware location-based data caching.

In embodiments, quantum computing system 17000 may include quantum computing resources, traditional computing resources, energy resources, human resources, robotic fleet resources, smart container fleet resources, I/O bandwidth, storage resources, It can be utilized for optimization of various system resources in one of the systems, including optimization of network bandwidth, attention resources, etc.

Quantum computing system 17000 may be implemented in a system of a system architecture similar to intelligent service 17034, where the full range of capabilities are available to or as part of any configured service. Composed quantum computing services may consist of subsets of these capabilities to perform certain predefined functions, create newly defined functions, or various combinations of the two.

Figure 161 illustrates quantum computing service request handling according to some embodiments of the present disclosure. Directed quantum computing request 17102 may come from one or more quantum-aware devices or stacks of devices, where the request refers to a known application consisting of specific quantum instance(s), quantum computing engine(s), or other quantum computing resources. and the data associated with the request may be preprocessed or otherwise optimized for use with quantum computing.

A generic quantum computing request 17104 may come from any system within a system of systems or configured services, where the requester has determined that quantum computing resources can provide additional value or other improved results. Improved results could also be offered by quantum computing services in relation to some forms of monitoring and analysis. For a typical quantum computing request 17104, the input data may not be structured or formatted as required for quantum computing.

In embodiments, external data requests 17106 may include any available data that may be needed to train new quantum instances. The sources of these requests can be public data, sensors, ERP systems, and many others.

Incoming operation requests and associated data use a standardized approach to identify one or more possible sets of known quantum instances, quantum computing engines, or other quantum computing resources that can be applied to perform the requested operation(s). can be analyzed. Potential existing sets may be identified in the quantum set library 17108.

In embodiments, quantum computing system 17000 includes quantum computing configuration service 17010. The quantum computing configuration service may operate alone or in conjunction with intelligence service 17034 to select the best available configuration using resource and priority analysis that also includes the requester's priorities. The quantum computing configuration service may provide a solution (yes) or determine that a new configuration is needed (no).

In one example, the requested set of quantum computing services may not exist in quantum set library 17108. In this example, one or more new quantum instances must be developed (trained) using the available data. For example, a quantum computing module for optimizing truck freight delivery in the United States may exist in quantum set library 17108. However, requester input identified a need to optimize shipping in Canada. In this case, quantum instance training could work with intelligence services 17034 to train a new instance for Canada using a variety of public data such as delivery schedules, speed limits, fuel mileage and costs, etc. In embodiments, alternative configurations may be developed with assistance from intelligence services 17034 to identify alternative ways to provide all or some of the requested quantum computing services until appropriate resources become available. there is. For example, a quantum/traditional hybrid model may be possible that provides the requested service but at a slower rate.

In embodiments, alternative configurations may be developed with assistance from intelligence services 17034 to identify alternative and possibly temporary methods for providing all or some of the requested quantum computing services. For example, a hybrid quantum/traditional model may be possible that provides the requested service but at a slower rate. This may also include feedback learning loops to adjust services in real time or improve stored library elements.

When a quantum computing configuration is identified and available, it is assigned and programmed for execution and delivery of one or more quantum states (solutions).

Biology-Based Systems For VCNs

The techniques described herein improve the ability of networks and systems to collect, transmit, and process large amounts of data, particularly data from sensors and other value chain data generators. These techniques include using the biological thalamus, a thalamus service that provides the equivalent of a neural system for filtering and relaying data. The awards service described herein receives large amounts of information and quickly prioritizes the information while delivering the most imported information so that limited transmission, processing, collection, and/or analysis resources are not overwhelmed by the volume of incoming information. You can rank them.

Additionally, the predictive model communication protocol (PMCP) is described herein. PMCP can be used to reduce the volume of transmitted data, especially when the data is predictable or generally predictable. PMCP may operate by transmitting predictive model parameters in place of some or all of the data values that would normally be transmitted by a sensor device or other data source. For example, a device implementing PMCP can continuously receive inputs (e.g., sensor data) and use the stream of sensor data to train a predictive model. Instead of transmitting sensor data, which can use up significant network and/or processing resources, PMCP devices can transmit model parameters that can be used by the receiving device to run predictive models to predict current and future sensor data. there is. Accordingly, the receiving device may have a prediction model of the sensor data without receiving the sensor data. In embodiments, if sensor data at a PMCP device begins to behave outside of expectations, model parameters can be retransmitted to the receiving device, which can update the prediction model and thereby obtain more accurate prediction data.

In some embodiments, quantum computers and/or predictive models may be used with the techniques described herein to optimize decision-making. Additionally, quantum coordination can be applied to allow different units to safely coordinate actions (eg, without the need for traditional communication mechanisms). Accordingly, the techniques described herein may use a combination of quantum capabilities and distributed biology-based decision-making capabilities distributed across devices in a value chain network. Additionally, these technologies are also efficient mechanisms to enable operational efficiency of coordinated value chain networks.

162 illustrates a set of input sensors streaming data from various sources across the Value Chain Network Awarding Service 18000 and the Value Chain Network and Systems (SOS) Control System 18002 along with their centrally managed data source 18004. do. The thalamus service 18000 filters input to the control system 18002 from various data sources 18004, ensuring that the control system is never overwhelmed by the total volume of information. In embodiments, awards service 18000 provides an information suppression mechanism for information flows within the value chain. This mechanism monitors all data streams and suppresses and/or filters irrelevant data streams by ensuring that the maximum data flows from all input sensors are always constrained.

In embodiments, award service 18000 may be a gateway for all communications responsive to the system's priorities among system control system 18002. A system of the system control system 18002 may determine to change the priority of data streamed from the city service 18000, for example, during a known fire in an isolated area, where the event may result in the majority of such data being abnormal ( You can instruct the vision service (18000) to continue providing flame sensor information despite the fact that it is unusual. The thalamus service 18000 may be an integral part of the overall system of system communication framework.

In embodiments, awards service 18000 includes an intake management system 18006. Intake management system 18006 receives multiple large data sets by converting them into data streams that are sized and organized for subsequent use by a central control system 18002 operating within one of the systems. and may be configured to process. For example, a robot may be used by a central control system 18002 (which may be mounted on the robot and/or on a separate device in communication with the robot) to identify the environment in real time and move through the environment. It may include vision and sensing systems. Intake management system 18006 may facilitate robotic decision making by parsing, filtering, sorting, or otherwise reducing the size and increasing the usefulness of multiple large data sets that would otherwise overwhelm central control system 18002. . In embodiments, the intake management system may include an intake controller 18008 that operates in conjunction with an intelligence service 18010 to evaluate incoming data and take action-based evaluation results. Assessments and actions may include the use of specific instruction sets received by award service 18000, for example, a set of specific compression and prioritization tools specified within the “Networking” library module. In another example, award service inputs may direct the use of specific filtering and suppression techniques. In a third example, award service inputs may specify data filtering associated with an area of interest, such as a specific type of financial transaction. The intake management system is also configured to recognize and manage data sets in a vectorized format, e.g., conforming to the Predictive Model Communication Protocol (PMCP) (discussed below), wherein the data sets are transmitted to the central control system 18002. , or alternatively, they can be broken down and processed separately. Intake management system 18006 may include a learning module that receives data from external sources to enable improvement and creation of application and data management library modules. In some cases, intake management system 18006 may request external data to augment existing data sets.

In some embodiments, SOS control system 18002 may instruct award service 18000 to change filtering to provide more input from a particular set of sources. This indication to provide more input is handled by award service 18000. For example, a presentation service may suppress other information flows to limit the overall data flow to a volume that the central control system can handle.

In embodiments, award service 18000 may operate by suppressing data based on several different factors, including zero or more default factors. For example, in some embodiments, default factors may include an “unusualness factor,” which may be a value that indicates the degree of divergence or divergence of the data from the expected data set. In embodiments, an abnormality factor is continuously monitored for all inputs or some of the inputs (eg, some of the input sensors).

In some embodiments, awards service 18000 may suppress data based on geospatial factors. Examples of geospatial factors include position data, motion data, acceleration data, vibration data, and/or any other data representing absolute or relative position, change in position over time, other derivatives or integrals of position over time, etc. Can contain data. Visualization service 18000 may recognize geospatial factors for some or all of the sensors and thus find unusual patterns in the data based on the geospatial context and suppress the data accordingly.

In some embodiments, awards service 18000 may suppress data based on temporal factors. For example, data may be temporally suppressed if the cadence of the data may be reduced such that the entire data stream is filtered to a level where it can be processed by the SOS control system 18002 and/or the central processing unit. .

In some embodiments, awards service 18000 may suppress data based on context factors. In embodiments, context-based filtering is a filtering event in which awards service 18000 recognizes some context-based events. Context-based events may include, for example, one or more notifications of abnormal behavior by other sensors or systems (which may result in temporary suppression of less sensitive data), one or more human inputs (e.g., (a human disabling a security alert can suppress previous focus on security alerts), one or more events triggered by other systems or sensors (e.g., resources can be diverted to security data collection, transmission, and analysis) automated security alerts that may result in the suppression of certain data in order to enable may result in the suppression of certain data until it is available), or any other context-based condition or event. In this context, filtering can suppress information flows that are not related to data from an event.

In embodiments, awards service 18000 may collect data from various data sources 18004, including analytics results 18018, databases 18020, sensors 18022, and/or reports 18024. can receive. For example, awards service 18000 may receive analysis results and/or reports from other analysis/processing/reporting devices with previously preprocessed sensor data or other data. Additionally or alternatively, awards service 18000 may receive data stored in database 18020 (e.g., historical data) in addition to current or historical data from sensors 18022. In embodiments, data may be received and/or generated from PMCP device interface 18052 (e.g., predictive models may generate future data).

In an embodiment, awards service 18000 may include a networking library 18014, a security library 18016, and/or any other libraries for interpreting various types of input data. Input from any of the data sources 18004 may be processed and/or interpreted based on the input. For example, awards service 18000 may parse, interpret, extract, and/or otherwise process network data (e.g., data received from networking sensors or devices, networking analytics, networking reports, network database data, etc.) You can use the networking library (18014) to do this. Similarly, awards service 18000 may parse, interpret, extract, and/or collect security data (e.g., data received from security sensors or devices, security analytics, security reports, security database data, etc.). You can use the security library 18016 to handle it in a different way. In embodiments, the intake data may also include an intake learning module that may use one or more artificial intelligence techniques to preprocess the data, use the data to create predictive models, predict future states of the data, etc. This can be processed using (18026). After processing using the Intake application library 18012 and/or the Intake learning module 18026, the data may be prepared for management by the Intake data management system 18028.

The intake data management system 18028 may use data to be prioritized 18030, formatted 18032, suppressed 18034, area focused 18036, filtered 18038, and/or combined (18038). You can process data by combining)(18040). Prioritization 18030 may involve prioritizing or otherwise assigning priorities (e.g., categories, numeric priority scores, etc.) so that limited resources can be allocated to the most important data. You can. For example, suppressing (18034) and/or filtering (18038) the least important data (e.g., data associated with the lowest priority score) to avoid overwhelming limited transmission, processing, and/or analysis resources. It may operate based on priorities to suppress or filter. Formatting 18032 may involve formatting data to allow for easier management, which may involve compressing or otherwise dropping certain portions of the data to reduce the use of transmission resources. decompressing data to reduce the use of resources (e.g., when bandwidth is sufficient and the data is important); decompressing data to emphasize or de-emphasize certain aspects; This may involve formatting, or otherwise adjusting the formatting. In embodiments, formatting 18032 may rely on prioritization 18030 such that more important data may be formatted to allow for more or better analysis, while less important data may be subject to the use of various resources. Can be formatted to reduce .

In embodiments, suppressing 18034 may involve reducing the amount of data, the number of destinations to which the data is transmitted, or otherwise utilizing limited resources of the data (e.g., bandwidth, processing, analysis, etc.). This may involve reducing use. In embodiments, suppressed data may be stored (e.g., in a database) and later processed (e.g., transmitted, processed). In embodiments, inhibition 18034 may be based on various factors as described above.

In embodiments, region focus 18036 may involve increasing attention made to specific high priority data. For example, during a security incident, security sensor data may be sent to additional destinations, processed using additional analytics, allowed additional bandwidth and processing power, etc. In embodiments, region focus 18036 may cause suppression or filtering of other data not associated with region focus 18036.

In embodiments, filtering 18038 involves ignoring, deleting, or otherwise removing data that is not important (e.g., does not match region focus 18036, is low priority, etc.) can do. In embodiments, data may initially be suppressed (e.g., reduced or stored for later), but conditions may be further changed to cause data to be filtered (e.g., deleted, ignored). there is. Accordingly, the intake data management system 18028 may allow gradual downgrade of data by first suppressing and later filtering the data depending on conditions.

In embodiments, combination 18040 may provide better analyses, generate new data, reduce the volume of data (e.g., by combining multiple data values into a single data value), (e.g., It may include combining various types of data to improve the quality of the data (for example, by averaging different sensor readings to obtain a more accurate average reading), etc. In some embodiments, lower priority data can be combined with other data to reduce requirements. Additionally or alternatively, higher priority data may be combined with other data to improve data quality.

In embodiments, the intake data management system may interface with intake controller 18008 and/or intelligence system 18042. Intelligent system 18042 may perform intake data management (e.g., prioritize data, format data, suppress data, select area focus, etc.) using various artificial intelligence techniques, for example. and/or assigning data to region focuses, filtering data, combining data, etc.), predicting results of intake data management, predicting future data values, etc. Additionally or alternatively, the intake controller 18008 and/or the intelligence system 18042 may control intake data management, artificial intelligence techniques (e.g., the parameters may be model parameters for an AI model) and/ Alternatively, the intake management system 18006 may operate according to configured parameters 18044 that may configure the operation of the intake management system 18006.

In embodiments, control system 18002 may, in some cases, provide quantum computing services 18046 that can provide quantum computing resources to more quickly process large amounts of data, use quantum models, etc. You can use it.

Control system 18002 may further include one or more data interfaces 18048 for receiving data from various data sources 18004 and transmitting data to various destinations (e.g., after managing intake data). In embodiments, control system 18002 may include analysis subsystems, various processing chips, or any other subsystem that can use managed data to make decisions, generate analyzes, or otherwise perform data operations. It may include other system subsystems 18050, such as: In embodiments, intelligence services 18010 may operate to route managed data to various other system subsystems 18050 or otherwise perform initial and/or final processing on the data.

In embodiments, the SOS control system 18002 may decide to override the thalamic filtering and focus on a different region for any particular reason. For example, during a security incident, the SOS control system may perform temporal filtering (usually can be deprioritized) and routed around. As another example, during routine inspections of equipment, sensor data that measures the operation of the equipment (e.g., vibration sensor data) is suppressed, even though the data appears to be within normal parameters and therefore can generally be suppressed or filtered. It may not work.

In an embodiment, control system 18002 may include a PMCP device interface 18052 that may be used to transmit and/or receive data using PMCP. Details of the PMCP device interface are further shown in second PMCP device interface 18060. In an embodiment, PMCP device interface 18052 may communicate with PMCP device interface 18060. PMCP device interface 18052 may have the same components as shown in PMCP device interface 18060.

In embodiments, a PMCP device interface may be used to convert data to a vectorized format prior to transmission. In these embodiments, the vector can be considered an example of a simple predictive model (e.g., a vector can indicate the amount of change and the direction of change for a data value, and thus the future state of the data value if the change continues). can be predicted). For example, the transformation of a long sequence of often similar data values into a vector, which represents the amount and direction of change, can imply the future state of the data values, allows communication of the data values to a smaller size and is essentially forward looking. (forward looking) makes both possible.

In embodiments, PMCP is a weighted moving average; Kalman filtering; exponential smoothing; ARMA (autoregressive moving average) (predictions depend on historical values of the variable being predicted and historical prediction errors); ARIMA (autoregressive integrated moving average) (ARMA for period-to-period changes in predicted variables); extrapolation; linear prediction; Trend estimation (forecasting a variable as a linear or polynomial function of time); growth curves (e.g., statistics); Various types of prediction models can be used to predict current and future data values, including recurrent neural network-based prediction.

Using the PMCP protocol, instead of traditional streams where individual data items are transmitted, vectors are communicated that indicate how the data is changing or what the predicted trends in the data are. The PMCP system can transmit actual model parameters to receiving units so that edge devices can apply vector-based predictive models to determine future states. For example, each automated device within the value chain network may be configured to train a regression model or neural network that continuously fits data streams to current input data. In some embodiments, automated devices utilizing a PMCP system may proactively react to events that actually occur, for example, rather than waiting for depletion of inventory for an item to occur. Continuing the example, a stateless automation device can react to predicted future states and make necessary adjustments, such as ordering more items.

In embodiments, a PMCP system enables communicating vectorized information along with model parameters that enable predictive models on the receiving end to predict probabilities of future values. Vectorized information can be transmitted and processed to determine multiple probability-based states. For example, motion vectors and model parameters for predicting future positions based on motion vectors may be transmitted using PMCP, and the received position may be configured to use the motion vectors to parameterize the prediction model (e.g., model parameters can be used as inputs to a model that determines the future locations of the item, which can generate probabilities that the item associated with the motion vector is at different locations. As another example, a PMCP system may support communicating vectorized sensor readings with model parameters that allow current and/or future sensor readings to be predicted. Applied in environments with a large number of sensors with different accuracies and reliability, the probabilistic vector-based mechanism of PMCP systems allows many, if not all, data streams to determine the current status of value chain items (e.g. goods, services, etc.). It can be used to create improved models representing states, past states, and likely future states. Approximation methods may include critical sampling, and the resulting prediction model may be a particle filter, condensation algorithm, Monte Carlo localization, or other suitable models.

In embodiments, the vector-based communication of a PMCP system allows devices and/or other systems to anticipate future security events. For example, a set of simple edge devices can be configured to run semi-autonomously using PMCP to generate and transmit model parameters based on locally sensed security data. In this example, edge devices can be configured to build a set of predictive models that show trends in the data. The parameters of this set of predictive models can be transmitted using a PMCP system. In this example, edge devices can be configured to build a set of predictive models that show trends in the data. The parameters of this set of predictive models can be transmitted using a PMCP system so that the secure data can be reconstructed and used to predict future states at the receiving device.

In embodiments, because abnormal events tend to cause one or more vectors to exhibit unusual patterns, security systems may generate and transmit vectors representing changes in state. In a security setting, detecting multiple simultaneous abnormal vectors may trigger escalation and response by, for example, a control tower or other systems within one of the systems. Additionally, one of the main areas of communication security issues concerns the protection of stored data, in vector-based systems data may not need to be stored (or can be stored on fewer devices), and thus the risk of data loss is high. eliminated or reduced.

In embodiments, PMCP data may be stored directly in a queryable database where the actual data is dynamically reorganized in response to queries. In some embodiments, PMCP data streams can be used to recreate fine-grained data, so that they become part of an Extract Transform and Load (ETL) process.

The PMCP device interface may include several modules including a transceiver module 18062, a modeling module 18064, a library module 18066, and a storage module 18068. The transceiver module may transmit various data from data source 18004 and/or to/from another PMCP device interface (e.g., PMCP device interface 18052) and/or to/from other components of the system that include a PMCP device interface. May include a data transceiver 18070 that can be used to transmit/receive data including PMCP data (e.g., vectors, model parameters, etc.). In embodiments, transceiver module 18062 may include an intelligence system 18072 that may use artificial intelligence techniques to assist with transmit and/or receive processing. For example, intelligent system 18072 may route various types of incoming and outgoing data, prioritize or de-prioritize incoming and/or incoming data from data source 18004 versus PMCP data, etc. The intelligence system 18072 may further include a PMCP controller 18074 that can understand the PMCP transmission, parse the PMCP data, and provide the received PMCP data to a modeling module for further operation.

Modeling module 18064 may be responsible for various operations in a transmit role and/or a receiver role. In the transmitting role, modeling module 18064 continuously receives data from various data sources 18004 (e.g., sensors 18022) and continuously creates and models models that predict future states of the incoming data. /Or it can be improved. The various models may be, for example, classification models, behavior analysis models, prediction models, data augmentation models, and/or any other types of models. Model parameters (e.g., neural network weights) from the generated/refined models can then be transmitted to receivers, which can perform classification, behavior analysis, prediction, etc. without having to access the data stream. Parameters can be used to perform augmentation, etc. Accordingly, in the receiver role, modeling module 18064 parameterizes various types of models using various parameters received from other PMCP device interfaces and then uses the parameterized models to provide data for further use by the receiving device. can be created.

In embodiments, the PMCP device interface may train and/or run a classification model 18076, which may be trained using data captured from a data source 18004 to generate various labels or classifications. For example, classification models can be used to output various states or conditions based on input data, including predicted future states or conditions. By sending classification model parameters to a receiving device using PMCP, the receiving device can also predict future states or conditions without having to receive input data from data sources 18004.

In embodiments, the PMCP device interface trains and trains behavior analysis models 18078 that can be trained using data captured from data sources 18004 to generate various behavior analyzes and future behavior data. /or can be executed. For example, behavior analysis models can be used to output analyzes of current or future actions likely to be taken by certain entities and/or whether actions are within normal conditions or abnormal. By transmitting behavior analysis model parameters to a receiving device using PMCP, the receiving device can also predict future actions and/or analyzes without having to receive input data from data sources 18004.

In embodiments, the PMCP device interface trains prediction models 18080 that can be trained using data streams captured from data sources 18004 to generate current and predicted data values for the data streams. and/or executable. For example, predictive models can be used to output current or future sensor readings based on data captured from sensors 18022. By sending prediction model parameters to a receiving device using PMCP, the receiving device can also predict sensor values without having to receive input data from sensors 18022 or other data sources 18004.

In embodiments, the PMCP device interface may train and/or execute augmented models 18082, which may be trained using data captured from data sources 18004 to generate augmented data streams. For example, augmented models can be used to generate interpolated or extrapolated values from data streams that may be missing data (e.g., due to network outages) or sensors from other nearby sensors. Based on the readings, predicted sensor readings for a sensor (e.g., a broken sensor) may be generated and data received from data sources 18004 may be otherwise augmented with additional data. By sending augmented model parameters to a receiving device using PMCP, the receiving device can also generate missing data, predicted data, or other augmented data without having to receive input data from data sources 18004. .

In embodiments, PMCP device interface 18060 may use a library module 18066 that contains one or more modules that may be used to assist with modeling and/or other operations. For example, networking module 18084 may be used to train network devices, network topologies, network digital twins, and various models, perform ETL operations as described in more detail below, or other such processing. It may contain various data about other network data that can be utilized to perform. As another example, security module 18086 can be used to train security devices, building layouts (e.g., for building security systems), maps, topologies, digital twins, vulnerabilities, and various models, as described below. may include various data regarding other secure data that may be utilized to perform ETL operations, as described in more detail herein, or to perform other such processing for security reasons. Various other specific modules may be provided to enable or support specific use cases.

In embodiments, storage module 18068 may provide various operations for processing data and/or storing data for storage. The ETL interface 18088 may be configured to perform exchange, transform, and load (ETL) operations to store data in the PMCP database 18090. PMCP database 18090 contains various data, including data received from data sources 18004 (e.g., so that historical data may be used to create/refine various models), as well as the models themselves, It can be used to store parameters, etc.

In embodiments, the awards service and PMCP may provide complementary technologies for managing large amounts of data. For example, PMCP can reduce bandwidth and storage requirements for working with large amounts of data because PMCP can only require transmission model parameters instead of transmitting bandwidth-intensive data streams. However, when processing a large number of data sensors or other data sources, the number of PMCP streams, number of models, etc. may still be too large to handle, so PMCP is sufficient to reduce the data to manageable levels. You may not. In such cases, the award service may act to prioritize, format, suppress, filter, or combine PMCP data streams to focus on the most important PMCP data streams at any given time. Several advantages are realized by combining technologies in this way. For example, large amounts of data may be collected, but PMCP may allow communication of model parameters to predict some or all of the data, and the award service may focus on the most important models and predictions at any given time. You can. Additionally, the use of PMCP allows the data to be inherently predictive and therefore forward looking, which, in combination with award services, can be used to track various actions (e.g. interventions, maintenance, purchase orders, supply Allows focusing on the most important data before potential problems arise that may require adjustments (adjustments, estimate adjustments, etc.).

163 illustrates the interaction of PMCP with various other components of intake controller 18008, intake management system 18006, and awards service 18000 in accordance with some embodiments of the present invention. In the illustrated embodiments, inputs may be received into intake controller 18008 from different sources. For example, the first source of data may be received from various sensors, external systems, process data, and other such data (which may be received from various data generators, data analysis systems, and other data outputs external to the awards service). 18102). Additionally or alternatively, the second source of data may include one or more pre-configured PMCP devices with position processing provided in 18104, which may include PMCP model parameters, vectorized data, or other PMCP data. there is.

An intake controller 18008 may collect the data and, at decision 18106, determine whether the data is PMCP data. If the data is not PMCP data, the intake controller 18008 can determine whether the data has been reduced. If the data has not been reduced, the data may be sent to an intake management system for processing (e.g., prioritizing, formatting, suppressing, region focusing, filtering, combining, etc., as discussed above). In other words, if the data has not already been reduced in some way (e.g., via PMCP or using other data reduction techniques), the data can be processed and potentially filtered, suppressed, or otherwise reduced. . Accordingly, the award service may provide one data reduction technique that can be used in addition to or as an alternative to other data reduction techniques, which may include PMCP.

If the data was not PMCP data as determined in 18108, or if the data was PMCP data as determined in 18106, intake controller 18110 may determine whether the awards service is acting as a PMCP consumer for the data. If so, the data may be sent to PMCP device interface 18052 for reception and processing (e.g., modeling, prediction, etc.). Otherwise, one or more ETL processes may be used to extract, transform, and load data from the 18114 into the PMCP database.

Whether the data is processed at 18114 by the PMCP device interface 18052 or using an ETL process, the resulting data may be provided to a system data consumer's downstream system for further processing at 18116.

PMCP and award service techniques can be used (together or separately) in a wide variety of embodiments. In embodiments where edge devices are configured with very limited capacities, additional edge communication devices may be added to convert data to PMCP format. For example, to protect distributed medical equipment from hacking attempts, many manufacturers will choose not to connect their devices to any kind of network. To overcome these limitations, medical equipment can be monitored using sensors such as cameras, sound monitors, voltage detectors for power usage, chemical sniffers, etc. Functional unit learning and other data techniques can be used to determine actual usage of medical equipment separate from network functional units, generate vectorized data therefrom, and/or transmit various model parameters using PMCP. At the receiving end, the award service may receive vectorized data and/or model parameters and determine whether PMCP data and/or other data received from other medical devices should be prioritized, filtered, suppressed, etc. You can use a variety of techniques to predict future states of medical equipment based on PMCP data, take various actions, and use some or all of the data to perform various analyses.

In some embodiments, communication within the value chain using vectorized data allows the value chain to have a consistent view of what likely future states are. These technologies allow future states to be communicated to the value chain, allowing value chain entities to respond in advance to future state requirements without requiring access to fine-grained data.

In some embodiments, the PMCP protocol transmits and receives relevant information regarding production levels and future trends in production (e.g., important or high priority information, as determined by the awards service) to various external entities. It can be used to: In some of these example embodiments, PMCP data feeds may be used for data obfuscation (e.g., communicating sensitive data as vectorized data and/or model parameters). For example, PMCP allows real-world contextual information about production levels to be shared with consumers, regulators, and other entities outside the value chain network without directly sharing sensitive data values. For example, when a customer chooses to purchase a new car, one or more value chain entities may be integrated into the selection process and determine (e.g., based on a predictive model) that there is an upcoming shortage of red paint. You can. In this case, the value chain entity can communicate PMCP data to be processed and used to show the impact of different choices on delivery times to the customer device, without providing sensitive data to the customer device or other external entities.

PMCP and vectorized data processing also enable simple data-informed interactive systems that users can apply without having to build incredibly complex big data engines. By way of example, an upstream manufacturer may have the very complex task of coordinating many downstream consumption points. Through the use of PMCP and/or award services, manufacturers can provide real information to consumers without the need to store detailed data and build complex models, which saves setting up large-scale systems to process large amounts of data, etc. You can request it.

In embodiments, edge device units may communicate via a PMCP system to show direction of movement and possibly future locations. For example, mobile robots can communicate likely future movement tracks. In embodiments involving large numbers of mobile robots, the award service may determine which robots need to be prioritized and closely monitored (e.g., because they are moving outside defined boundaries). , because they behave in unpredictable ways, etc.).

In embodiments, the PMCP system and/or vision system enables visual representations of vector-based data (e.g., via a user interface), including highlighting regions of interest without the need to process enormous amounts of data. . The visual representation allows the display of many monitored vector inputs. The user interface can then display information about key items of interest, specifically vectors representing areas of abnormal or difficult movement. This mechanism allows sophisticated models built on edge device edge nodes to feed end-user communications in a visually informative manner.

As can be appreciated, the functional units generate a constant stream of “boring” data (eg, data that does not change, changes slightly, or changes very predictably). By shifting from data generation to problem monitoring, problems in logistics modules can be highlighted without delving into micro-granular data. In embodiments, PMCP device interfaces may continuously create and/or improve predictive models that predict future states. In the context of maintenance, improvements to parameters in a predictive model are within and themselves predictors of changes in operating parameters, potentially indicating a need for maintenance. Moreover, communication of operating parameters for multiple devices may be handled by the award service such that data for normally functioning devices may be filtered or suppressed until conditions change.

In embodiments, functional areas are not designed to always be connected to the value chain network, but instead allow external devices to virtually monitor the devices, thereby allowing functional areas that do not allow connectivity to access information flow in value chain products. can be a part This concept allows functional areas with limited connectivity to be monitored effectively by embellishing the data stream with vectorized monitoring information. Placing automated devices near functional devices with limited or no connectivity allows information to be captured from the devices without connectivity requirements. There is also the possibility of adding training data capture functional units for these unconnected or limitedly connected functional areas. These training data capture functional units are typically quite expensive and can provide high quality monitoring data that is used as input to a near edge device monitoring device to provide data for supervised learning algorithms.

Often, value chain network locations suffer from electrical interference, causing fundamental problems in communication. Traditional approaches to streaming all fine-grained data rely on the integrity of the data stream. For example, if an edge device goes offline for 10 minutes, streaming data and its information will be lost. With vectorized communication, an offline unit can continue to refine its predictive model until the moment it reconnects, allowing the updated model to be transmitted over the PMCP system.

In embodiments, value chain network systems and devices may be based on the PMCP protocol. For example, value chain network cameras and vision systems (e.g., liquid lens systems), user devices, sensors, robots, smart containers, etc. may use PMCP and/or vector-based communications. For example, by using vector-based cameras, only information about the movement of items is transmitted. This reduces data volume and filters information about items that are static by their nature, showing only changes in images and focusing data communication on the elements of change. The overall shift from communication to communication of change is similar to how human visual processes function, where static items are not even transmitted to higher levels of the brain.

Radio frequency identification allows vast quantities of mobile tags (for example, cargo RFID tags for cargo transported by smart containers) to be tracked in real time in value chain networks. In embodiments, the movement of tags may be communicated as vector information via the PMCP protocol, as this form of communication is naturally suited to handling information regarding the location of tags within value chain products. Adding the ability to show the future state of a location using predictive models that can use paths of previous movement means that units consuming data streams from value chain products can consume information about the likely future state of value chain products. This allows changing the basic communication mechanism. In embodiments, each tagged item may be represented as a probability-based location matrix representing the probability that the tagged item is likely to be at that location in space. The communication of movement illustrates the transformation of the position probability matrix into a new probability set. This probabilistic location overview provides consistent modeling of areas of possible intersection of mobile units and allows refinement of the probabilistic view of the locations of items within the value chain network. Moving to a vector-based probability matrix allows units to continuously handle the inherent uncertainty in the measurement of the state of value chain network items, entities, etc. In embodiments, status includes, but is not limited to, location, temperature, movement, and power consumption.

In embodiments, a persistent connection is not required for continuous monitoring of sensor inputs in a PMCP-based communication system. For example, a mobile robotic device with multiple sensors can continue to build models and predictions of data streams while disconnected from the network, and upon reconnection, updated models are communicated. Additionally, other systems or devices using input from the monitored system or device can maintain a probabilistic understanding of the states of value chain commodities by applying the best known, typically last communicated, vector predictions.

Energy systems and processes

This disclosure relates to energy systems and processes. In example embodiments, there are energy systems and processes (also referred to as energy systems, which may be or include, e.g., energy systems, processes, modules, services, platforms, etc.). In an example embodiment, as shown in Figure 164, an energy system (e.g., a value chain network management platform) that can interact with any of the systems, subsystems, components, processes, platforms, etc. described in this disclosure (e.g., a value chain network management platform) 19000) exists. In some examples, energy system 19000 may be a separate system external to the systems, subsystems, components, processes, platforms, etc. in this disclosure. In other examples, energy system 19000 may be integrated with any one of the systems, subsystems, components, processes, platforms, etc. of this disclosure. For example, as shown in FIG. 164, energy system 19000 includes a value chain network entity 652 and a data handling layer 608 (e.g., value chain management platform 604, adaptive intelligence system ( 614), monitoring system 808, data storage system 624, interface 702, connectivity facility 642, and “process and application output and results” 1040. Energy system 19000 (e.g., including at least one energy process) may utilize an energy model. Energy system 19000 may be part of a group of value chain building blocks that may be combined with various other processes and/or systems within an enterprise control tower. In example embodiments, energy system 19000 may provide modular, adaptive resource package technology (e.g., with energy, energy computing, and/or energy networking processes). For example, energy system 19000 may relate to energy storage on a modular level (eg, use of a modular energy storage device) across a network of energy storage systems and devices. Energy Systems (19000) can address a variety of needs for power management across communities, businesses/companies, organizations, colleges/universities, and more. This can be achieved, for example, by modularization of power storage. In example embodiments, where there may be a need to focus on limited power resources and renewable energy, optimization and modularization of power storage may solve these problems.

In example embodiments, energy system 19000 may include and/or utilize any one or more of the following technologies, systems and/or processes: three-dimensional (3D) printing of batteries, battery energy storage system (BESS) ), various battery types, coordination processes, distributed energy grids, energy prices, energy storage technologies, energy-as-a-service (energy-as-a-service), e.g. ), energy-related sectors and transactions, machine learning (ML) and/or artificial intelligence (AI) for energy optimization, ML/AI for automation, decentralized networks (e.g. networks of energy production, storage and delivery systems) ML/AI for matching energy use/demand to energy production via ), quantum, renewable energy (e.g. renewable energy kits), technologies for slicing (e.g. systems and/or processes for use, etc.

In example embodiments, energy system 19000 may include energy storage technology. In some embodiments, energy storage technology may include one or more types of batteries. For example, batteries include lithium-ion batteries, flexible batteries, structural batteries, and solid-state batteries (e.g., technological advances have led to the emergence or re-emergence of solid-state batteries in some cases as commercial alternatives). may result in), and/or flow batteries.

In exemplary embodiments, new materials and manufacturing methods may lead to the introduction of flexible batteries (e.g., flexible primary and secondary cells), and a pipeline of new possibilities. Use case or product driven conformance and compliance capabilities can provide new parameters for optimizing product design. Flexible batteries can be designed as an integrated part of a product rather than as an add-on module and can be used in clothing and other wearable electronic devices, medical devices, drug delivery systems, micro IoT devices, flexible batteries, etc. flexible electronic devices incorporating both flexible circuits and batteries, etc.

In an exemplary embodiment, a structural battery may use carbon fiber as the cathode and lithium iron phosphate-coated aluminum foil as the anode. Trade-offs between battery weight and product performance could eliminate or limit battery-powered product categories. Technological advances may provide opportunities for designs that integrate at least one structural battery into the product itself. Batteries using carbon nanotube electrodes as structural elements can provide design flexibility and opportunities for overall weight reduction, for example as part of the hull of a ship. Structural storage elements can support battery-powered supplies that can integrate transaction capabilities, and possibly integrated systems for grid-independent infrastructure, vehicles, devices, etc. Exemplary uses may include sidewalks, roads, airports, etc. (public or private). Geometry, cost, structural requirements, design life, thermal management, power and energy, or a combination of these characteristics can be automated as part of the design and value chain consideration process.

In example embodiments, flow batteries can be used to separate energy and power. For example, using a type of flow battery that separates energy and power can provide some unique design and integration opportunities, such as purpose-built buildings and infrastructure. These flow battery systems could provide a short-term alternative to lithium-ion batteries. New chemistries, such as organic formulations, could lead to easier use of abundant and less corrosive electrolytes, which could make this technology less expensive.

In example embodiments, energy storage technology may include smart batteries. A smart battery may be a smart battery with a Battery Management System (BMS) and other functions down to the cell-level. A battery-level BMS can be used to manage charging, discharging, voltage balancing, etc. In other examples, smart batteries may be smart batteries with cell level monitoring and data streams. In other examples, smart batteries may be smart batteries with cell-level distributed energy management. In other examples, smart batteries may be smart batteries with energy management on a chip (e.g., chipset) for cell level or system level control. In an example embodiment, automated battery assembly monitoring, maintenance, and performance management may be simpler when the cells monitor themselves, freeing assembly control to perform higher level operations. For example, BMS and other functions down to the cell level can manage charging, discharging, voltage balancing, etc. In example embodiments, smart batteries may provide cell level monitoring and data streams, cell level distributed energy management, and/or energy management on a chip for the cell or system level. In example embodiments, on-chip energy management may provide cell or system level control. For example, a chip could be used to electrically switch cell connections to optimize balancing of power and energy requirements. Individual cells can request their own replacement or go offline. In example embodiments, smart batteries may utilize quantum computing for design or real-time operating temperature optimization, including automated design and system control. Smart batteries can also provide vibration control, such as controlled vibration to manage dendrites and improve battery life for lithium-ion, zinc, etc.

In example embodiments, energy storage technology may include various control and/or management functions. For example, energy storage technologies can manage dendrites (e.g., for lithium ion, zinc, etc.) and provide controlled oscillations to improve battery life. Energy storage technologies may provide battery product lifecycle management (e.g., battery product lifecycle management system) and/or battery management and control (e.g., batteries with wireless power and control).

In example embodiments, battery product lifecycle management can address supply chain optimization concerns for primary and intermediate battery materials and related opportunities. Battery product lifecycle management may involve more vertically integrated operations including integration of processes, battery production and co-location of products. In example embodiments, battery product lifecycle management may be used with battery manufacturing including material use optimized with value chain network (VCN) modeling and/or battery manufacturing with 3D printed materials and processes. In example embodiments, battery product lifecycle management may include data collection, management, and analysis that may incorporate testing and tracking of battery “cells” and other sub-components for VCN optimization. Battery product lifecycle management can also be used with battery recycling and reuse (e.g., lithium and/or cobalt recycled materials), as well as with battery disposal, carbon footprint management, etc.

In example embodiments, battery management and control, such as wireless power and control, may include wireless technologies for all levels of battery implementation and control. This can provide design and operational flexibility, such as charge/discharge control, real-time power/energy configurations, and/or operational notifications down to the cell level. In example embodiments, wireless power and control supports simplified system integration, system and software standardization, integrated higher-level power dispatch and control systems, and/or chip-level integrated circuits and power management. It may include BMS software to do this.

In other examples, energy storage technology may utilize battery-powered/grid-independent infrastructure. In example embodiments, structural storage elements can support integrated systems for battery-powered, possibly grid-independent infrastructure, vehicles, devices, etc., that can integrate transaction capabilities. Examples may include sidewalks, roads, airports, etc. (public or private infrastructure). In example embodiments, geometry, cost, structural requirements, design life, thermal management, power and energy, or a combination thereof may be automated as part of the design and value chain consideration process.

In example embodiments, energy storage technology may utilize high-performance electrodes and/or high-performance separators. According to example embodiments, high-performance electrodes may include graphene and/or nanotubes. These high-performance electrodes can allow for faster charge and discharge, fewer thermal management issues, and associated safety, cycle life, and other performance improvements. Electrode improvements can benefit almost all battery types, and they can represent design opportunities for more and higher performance product implementations. Some advances include various carbon compositions such as graphene and nanotubes, and other materials that can increase active surface area, provide better manufacturability, lower resistance, longer lifetime, etc. Emphasizes improvements in the same electrode. In example embodiments, high-performance separators can be used with some example battery technologies. For example, for some battery technologies, particularly flow batteries and fuel cells, separators can be a key element in allowing charge transfer without direct mixing of anolyte and catholyte components. Improved separators could result in more efficient battery operation, lower costs, and wider deployment.

In example embodiments, energy storage technology may utilize organic flow battery electrolytes and/or polymer lithium-ion chemistries. For example, energy storage technology may include batteries with organic flow battery electrolytes. There may be a series of incremental improvements to organic flow battery electrolytes that can be rolled into existing and improved infrastructure. Latest research can focus on low-cost and environmentally friendly options. Examples may include organic flow battery electrolytes, polymeric lithium-ion chemistries, etc. In example embodiments, energy storage technology may include batteries comprising polymeric lithium-ion chemistries.

In example embodiments, energy storage technology may utilize wave energy (e.g., a system for storing wave energy) and/or thermal energy (e.g., a system for storing thermal energy). For example, ocean and geothermal open and closed systems can provide interesting arrangements. According to one embodiment, energy storage technology may provide gravity energy storage. For example, gravity energy storage technology can be integrated into building and infrastructure projects and managed as part of an integrated energy management system. In exemplary embodiments, the energy storage technology includes carbon particles that generate an electric current by interacting with a surrounding organic solvent (e.g., energy having carbon particles capable of generating an electric current by interacting with a surrounding organic solvent). A system for generating a) can be provided.

In example embodiments, the energy system 19000 may include various types of batteries. These battery types may include zinc battery types, nickel battery types, and/or cobalt battery types.

In example embodiments, energy system 19000 may include systems and/or processes to provide battery energy storage. This could be about battery energy storage systems (BESS). Some example BESS technologies with existing deployments and other near future possibilities could be hydrogen fuel cells and various types of flow batteries (eg, vanadium-based batteries). The vanadium-based battery may be, for example, a vanadium redox battery (VRB), a type of rechargeable flow battery (e.g., known as vanadium flow batter (VFB) or vanadium redox flow battery (VRFB)). Other examples of BESS technologies may include pumped hydraulic power, gravity, heat, tidal currents, and waves.

In example embodiments, BESS may be integrated with a building energy management system. In the United States, the Standard for the Installation of Fixed Energy Storage Systems provides cover for lithium-ion and other energy storage deployments in commercial buildings with shared tenants (e.g., New York) and in densely populated areas. of Stationary Energy Storage Systems)”. There may be some drawbacks to using lithium-ion technology, such as safety, mining of raw materials, recycling, cycle life, total cost of ownership, quality control, temperature operating performance limitations, energy density, etc. Standards, shortcomings and other factors may spur various industry convergences, such as BESS integration with Building Energy Management (BEM) systems.

In one embodiment, the BESS may be a floating battery-based BESS. Flow batteries may be the next closest commercial-scale large-scale BESS. Flow battery systems can use much of the same power, control and data infrastructure used with lithium-ion deployments. Flow battery systems may not pose a fire risk compared to lithium-ion systems, have fewer restrictions on charge/discharge cycles, and may have lower cost of ownership over a 20-30 year period. A typical flow battery electrolyte can be easily recycled or reused, and in some financing models, the flow battery electrolyte can be leased to reduce implementation costs. Technological advances associated with lithium-ion batteries, such as power electronics, controllers, electrodes, separators, and in some cases electrolytes, can also provide performance and cost improvements for flow battery systems.

In example embodiments, energy system 19000 may include systems and/or processes for providing 3D printing of batteries. This can create various types of 3d printed batteries by utilizing a 3d printer to print the batteries. Supply chain optimization for primary and intermediate battery materials may be a concern and opportunity that may relate to more vertically integrated operations that may involve integration of processes involving battery production and juxtaposition of products. In example embodiments, battery manufacturing may include material use optimized with value chain network (VCN) modeling. In other examples, battery manufacturing may include 3d printing materials and processes. In example embodiments, data collection, management, and analysis may incorporate testing and tracking of battery “cells” and other sub-components for VCN optimization. There may be systems and/or processes involving battery disposal, carbon footprint management, etc. There may be other systems and/or processes that can provide for battery recycling and reuse (e.g., lithium-ion battery recycling and reuse, such as recycling and reuse of lithium).

In example embodiments, energy system 19000 may include renewable energy technology (e.g., a renewable energy kit). This may relate to a system for generating, storing and/or using renewable energy (e.g. a renewable energy kit/in-a-box).

In example embodiments, energy provider(s) may include various options such as purchasers, servicers, self-generated, private/public, and/or mixed combinations. In example embodiments, energy source(s) may include various options such as solar, wind, batteries, thermal, gravity, waves, and/or grid.

In example embodiments, energy system 19000 may include decentralized energy grids. These distributed energy grids may include safety systems for distributed virtual grids. In some examples, distributed energy grids may include control systems for distributed virtual grids.

In other examples, distributed energy grids may allow transactions between end users. In example embodiments, a distributed energy grid may allow for disparate energy generation sources (e.g., solar, hydro, wind) and transactions between end users for generated and unneeded excess energy. can be allowed. Home energy assets can be tokenized and bought or sold on a decentralized marketplace. Transactions may be between users belonging to the same geographic location or different locations and may take price differentials into account. Excess energy can be supplied to the energy grid through smart contracts. Energy data (e.g., individual or aggregated data within a cluster) can be monetized by selling usage data.

In example embodiments, energy system 19000 may include systems and/or processes for serving energy-related sectors and transactions (e.g., energy transactions). “Energy-related sectors and transactions” may be used throughout this disclosure to refer to and incorporate such systems and/or processes for providing energy-related sectors and transactions. For example, energy system 19000 may include an energy trading system to facilitate energy transactions between parties. Energy-related sectors and transactions may provide for local and regional energy arbitration (e.g., using local and/or regional energy arbitration systems). Energy-related sectors and transactions may also provide local and regional energy management (e.g., using local and/or regional energy management systems). In some examples, energy-related sectors and transactions may provide an energy data marketplace (eg, using the energy data marketplace for individual or aggregated monetization of energy data). In other examples, energy-related sectors and transactions may include kiosks and/or microservices for energy in remote or scarce areas. In example embodiments, energy-related sectors and transactions may include a personal carbon usage monitoring and management system (eg, a system for monitoring and managing personal carbon usage). In other examples, energy-related sectors and transactions may include a corporate carbon usage monitoring and management system (e.g., a system for monitoring and managing corporate carbon usage and/or energy carbon usage). Energy-related sectors and transactions could include solar pumps and/or battery systems for crop irrigation that could support smart contracts. In some examples, energy-related sectors and transactions may use automated financing/payments/insurance mechanisms and/or smart contracts that can support private energy infrastructure investments (e.g., private energy infrastructure investments). may provide a system for automating financing, payments, insurance mechanisms, and/or smart contracts to support investments. In example embodiments, energy-related sectors and transactions may include the Gaming Engine smart contract energy management platform. The Gaming Engine smart contract energy management platform may be configured to enable energy management, energy visualization, modeling energy options, and/or smart contract execution.

In example embodiments, energy-related sectors and transactions may include energy ownership concepts and energy transactions between parties that may involve markets outside the utility penumbra, local and regional energy arbitration, and /or may include local and regional energy management. In other examples, energy-related sectors and transactions may involve monetization of data, such as personal or aggregated data. In example embodiments, energy-related sectors and transactions may be utilized in remote or low-service areas (eg, using kiosks/microservices). In other example embodiments, energy-related sectors and transactions may provide for personal and commercial lifecycle carbon monitoring and management. This may provide optimization of energy use mix, real-time cost offsets based on time of day, incentives, local regulations, etc., and/or may be part of a private value chain. In example embodiments, energy-related sectors and transactions may include integrated purpose-built systems (e.g., a solar-powered pump/battery system for crop irrigation supporting smart contracts). In other example embodiments, energy-related sectors and transactions may provide automated financing, payment, insurance mechanisms, and associated smart contracts that support private energy infrastructure investments. Energy-related sectors and transactions may also include Gaming Engine smart contract energy management platforms. The combined technologies of distributed hubs such as data centers, communications, power generation, storage, and dispatch can create multiple complex optimization scenarios. The gaming engine can be used for energy management, visualization, and contract execution. This may be an embedded or standalone platform, and may be license or subscription based. Applications of gaming engines include multi-tenant buildings, residential energy purchasing, residential use decisions, residential visualization, gaming engines embedded in products, and/or model energy options using gaming engines with smart contracts. It may be possible, but may not be limited to this.

Energy-related sectors and transactions may provide energy management (e.g., using energy management systems). In example embodiments, energy-related sectors and transactions may provide integration of multiple energy sources for storage and dispatch (e.g., using an energy management system). Energy-related sectors and transactions can also provide deployable, integrated and modular energy storage systems that incorporate interchangeability. For example, an energy management system may include an integrated modular energy storage system incorporating interchangeability.

In example embodiments, energy management systems may encourage broader and accelerated adoption as new technologies, lower costs, advanced regulations, blockchain distributed ledgers, electric vehicle integration, and energy management applications are deployed. can provide energy management. An energy management system may provide integration of multiple energy sources for storage and dispatch (eg, collocated or otherwise). Additionally, the energy management system may provide deployable integrated modular energy storage systems that may incorporate interchangeability (eg, a family of components similar to those associated with power tools). Energy management systems may include and/or be used in conjunction with commercial and building management systems. With the adoption of electric vehicles and large-scale BESS, energy management systems can provide smaller-scale distributed and separate storage/plus opportunities. Many small-scale land/grid connected systems can be used to address the cost of transporting and maintaining fuel oil for diesel generators in remote locations such as islands, mining sites, etc. Energy management systems provide packaged systems, or parts of packaged systems, for a broader set of customers that may include integration of various generating assets (e.g., wind, solar, diesel), storage, monitoring, and control. can do. In example embodiments, the energy management system may include and/or be integrated with home/community systems. For example, integrated control, energy management, trading, and market-enabling technologies for commercial and residential multi-tenant facilities could include electric vehicles and their batteries as well as stationary batteries, which could become more sophisticated. . In example embodiments, the energy management system may include new batteries for residential and commercial storage (eg, using smart batteries). In example embodiments, an energy management system may provide energy service contracts. For example, energy services contracts may involve an independent service industry associated with the installation and operation of residential solar systems that can be expanded to include a variety of integrated energy storage and service options. There may be opportunities for a wide range of contracted services.

In example embodiments, energy-related sectors and transactions may include a platform for dynamic allocation of distributed data center resources. For example, the platform may include a system for allocating resources based on energy cost, environmental impact, transaction volume, transaction type, and/or transaction priority. Energy-related sectors and transactions may also include integrated edge-based systems capable of generating and/or storing energy. In example embodiments, distributed and agile data centers may result in new infrastructure changes and strategies, such as integration of on-premises, co-location, cloud, and edge delivery options. The focus can be on workload placement. In example embodiments, a platform for dynamic allocation of distributed data center resources based on energy cost, environmental impact (e.g., legislation indicates movement in this area), transaction volume, type and priority of transactions, etc. So you can allocate resources. In example embodiments, integrated edge-based systems may be systems that generate energy, store energy, and/or provide data center-like services along with other energy management capabilities. These other energy management capabilities may include home systems, systems associated with one or more products with integrated intelligence, systems that may be available for a subscription fee from a service aggregator, integrated 5G communications infrastructure planning, and/or home May include and/or relate to underlying cryptocurrency operations.

In some example embodiments, energy-related sectors and transactions may provide analysis of land use costs. For example, analysis of land use costs may be accomplished by a system for analyzing the costs of renewable deployments, where the costs may be based at least in part on environmental, regulatory and/or zoning factors. . Land use can be part of the energy value chain, and environmental, regulatory, zoning and other local issues can be a cost to renewable deployment. This analysis can be used with a variety of examples, including floating wind and solar projects.

Energy-related sectors and transactions can also provide personal energy management. In example embodiments, personal energy management may be provided by a personal energy management system to manage personal energy use, storage, and/or generation. Inputs include personal energy assets and accounts, personal preferences (e.g. carbon footprint, budget), real-time data (e.g. hourly prices, cloud forecasts, wind forecasts), infrastructure (e.g. , local rules, interconnections), asset pricing models (e.g., depreciation, operating costs). This can provide time-sliced automated private microgrid control (e.g., active management, setup and forget).

In example embodiments, energy system 19000 may include regulation features. For example, such coordination features may include a system for coordinating energy demand across multiple distributed energy production, storage, and/or delivery systems. In example embodiments, coordination features may include coordination of energy demand across multiple distributed and partially isolated energy production, storage, and/or transmission systems.

In example embodiments, energy system 19000 may include systems and/or processes for providing energy pricing. Energy pricing may include pricing mechanisms that incorporate security, reliability, type-slicing, and/or time-slicing into the price matrix. These pricing mechanisms can be used with energy value chain networks, such as decentralized networks of production, storage, and delivery systems instead of centralized grids.

In example embodiments, energy system 19000 may include ML/AI for automation, which may relate to automation of energy transactions and/or energy management. In some examples, ML/AI for automation may include ML and/or AI systems for smart contract tracking (e.g., smart contract management) and/or price energy production on a blockchain system.

In other examples, ML/AI for automation may include ML and/or AI systems configured for automation of energy management in the supply chain (e.g., automating renewable energy management in the supply chain). For example, the use of renewable energy in the supply chain (e.g. in a factory or distribution center) will involve AI capabilities to track the supply and use of energy to ensure there is a sufficient amount of energy stored for daily use. You can. These ML and/or AI systems can monitor demand over the course of any period of time (e.g., hours, days, weeks, months, years). Fluctuations in demand can be anticipated by ML and/or AI systems, so that if demand reaches a threshold where new additional sources are needed, ML and/or AI systems can detect these before they become a problem for the supply chain/value chain. Demand can be predicted. Renewable energy can vary depending on location and region and may include solar, wind, water, geothermal, etc. ML and/or AI systems can also be used to make energy use more efficient. Similar to demand tracking, usage can be similarly tracked in relation to energy use to determine where inefficiencies may be realized, and then ML and/or AI systems can then inform the network about the supply chain/value chain. Suggestions can be made for adjustments in energy use across the globe.

In example embodiments, energy system 19000 may include ML and/or AI for energy optimization. ML and/or AI for energy optimization may further include and/or utilize the following systems and/or processes: ML and/or AI system configured to optimize the safety of lithium-ion batteries, lithium-ion batteries ML and/or AI system configured to optimize the cost of lithium-ion batteries, ML and/or AI system configured to optimize recycling characteristics of lithium-ion batteries, and/or ML and/or AI system configured to optimize food and energy production and storage .

In example embodiments, ML and/or AI systems for optimizing food and energy production may include a distributed food production value chain network. This may relate to food (e.g., heavy and therefore relatively expensive to transport based on energy demands) and plants (e.g., growing plants that may require relatively high energy densities). Food supply chains may be highly optimized and effective in the advanced industrialized world, but they can also be vulnerable to disruption, and for many foods (e.g. produce and meat), enormous amounts of transport are required to transport items that may consist primarily of water. Energy may be consumed. Food production can be energy intensive and in some instances requires a specific type of energy (e.g., a given set of spectral characteristics to promote plant growth over time). In some examples, a localized food supply chain can be managed simultaneously, locations can be provisioned, and the utilization of small-scale energy production entities, energy storage entities, delivery systems, and food production systems can be robust, efficient, and efficient. and can provide a food energy value chain network. In example embodiments, the platform may be configured to within a target budget (e.g., by time-shifting) to produce desired results (e.g., plant growth to meet forecast demand). Robotic process automation can be integrated to provision and deliver the desired mix of energy (e.g., for environmental goals). This may be achieved by DPANN techniques, quantum computing, and/or other optimization techniques in this disclosure. Demand-side forecasting can be applied to consumer demand for food and energy (e.g., based on a wide variety of IoT and crowd-sourced data). This may involve the aggregation of demand by robotic agents to a point that can justify the provisioning of production entities (e.g. a mix of infrastructure and food). Some goods can be produced with a minimal footprint (as in vertical farms), while others may require more land, but in either case, there is a need for production, storage and delivery systems for energy and food. Land use can be optimized by the system, such as scheduling temporary locations and considering individual parameters (e.g., energy requirements for food storage parameters).

In example embodiments, ML and/or AI for energy optimization may include a ML and/or AI system for optimizing energy utilization (e.g., for a specific location, time window, and application). . For example, large volumes of source production may be timed to meet close to the point of use (e.g., long-distance operations may not be a cost-effective option) or within a specific window of use (e.g., just before the asphalt is used). may require additional energy to keep the material flexible, deposit the material, and configure the material to prepare it for use. Field applications also require several energy consuming devices (e.g. spreaders, rollers, compressors) and materials (e.g. edge sealer, topcoat sealer, line striping). You can do this. In example embodiments, distributed energy systems are configured close to target areas for asphalt use that may require resurfacing (e.g., new developments, urban areas, or roads such as stretches of highways). It can be. With a flexible approach to energy and storage, construction crews can bring along the energy they need (or exceed demand).

In example embodiments, ML and/or AI for energy optimization may include and/or utilize the following systems and/or processes: ML and/or AI system configured for optimization of power grids, design optimization, ML and/or AI systems configured for (e.g. configured for design optimization of batteries), and/or ML and/or configured for real-time operating temperature optimization (e.g. configured for real-time operating temperature optimization of batteries) Or AI systems. In example embodiments, quantum computing for design or real-time operating temperature optimization may include automated design and system configuration control.

In example embodiments, ML and/or AI for energy optimization may include and/or utilize the following systems and/or processes: ML and/or AI systems for optimizing battery disposal and/or battery recycling; or ML and/or AI systems configured to optimize reuse. In example embodiments, these ML and/or AI systems may be involved in more vertically integrated operations, including integration of processes and co-location of products and battery production. Addresses concerns surrounding supply chain optimization for intermediate battery materials. For example, these ML and/or AI systems could address the needs of battery manufacturing, which could include material use optimized with VCN modeling. These ML and/or AI systems can be used for battery manufacturing using 3D printing materials and processes. These ML and/or AI systems may also be used with data collection, management, and analysis that may incorporate testing and tracking of battery “cells” and other sub-components for VCN optimization. ML and/or AI systems to optimize battery disposal can provide battery disposal, carbon footprint management, and more. ML and/or AI systems for optimizing battery recycling or reuse can provide battery recycling and reuse, such as the supply of lithium or cobalt needed for lithium-ion batteries.

In example embodiments, ML and/or AI for energy optimization may include a ML and/or AI system for optimization of energy use mix. In other example embodiments, ML and/or AI for energy optimization may be used to optimize the mix of energy sources and storage elements (e.g., involving the process of production, storage, use, and/or utilization of delivery systems). It may include ML and/or AI systems configured to optimize the production, storage, and utilization of

In other example embodiments, ML and/or AI for energy optimization may be implemented in systems and/or processes for providing energy cost optimization across decentralized commerce models (e.g., decentralized commerce models). ML and/or AI systems configured to optimize energy costs across models). These ML and/or AI systems include several manufacturing locations, multiple types of routes, and multiple types of transportation (e.g., third-party logistics (3PL), fourth-party logistics (4PL), super It can be used with VCNs that use grid logistics (super grid logistics, logistics markets). These ML and/or AI systems can optimize decisions about where/when to manufacture and how/when to ship, based on real-time and predicted energy costs as input. This will be achieved by monitoring all entities across the VCN and analyzing different variables to ensure that customer demands are being met while keeping energy costs to a minimum (e.g., using AI, predictive analytics, and quantum). You can. The same can be applied to multi-tenancy facilities to provide customers with the opportunity to optimize their energy usage costs. Additionally, predicting and securing future energy demand can be based on expected customer demand.

In example embodiments, energy system 19000 may be configured to match energy utilization and/or demand to energy production across a distributed network (e.g., a network of energy production, storage, and delivery systems). May include AI systems.

In example embodiments, energy system 19000 may include quantum features. For example, these quantum features include quantum for optimizing energy utilization for a location, time, and/or application (e.g., a quantum computing system for optimizing energy utilization for a specific location, time window, and application). can do. In example embodiments, a quantum computing system can address the production of large quantities of sources that can be timed to meet a specific window of use close to the point of use (e.g., just before the asphalt is used) and (e.g., Long hauls may not be a cost-effective option), and field applications may require additional energy to keep the material flexible, deposit the material, and configure the material to prepare it for use. Field applications may also require multiple energy consuming devices (e.g., spreaders, rollers, compactors) and materials (e.g., edge sealers, topcoat sealers, line striping). In example embodiments, distributed energy systems are configured in close proximity to target areas for asphalt use (e.g., new developments, urban areas, or roads such as stretches of highways) that may require resurfacing. It can be. With a flexible approach to energy and storage, construction teams can bring together the energy they need (or exceed their needs).

In example embodiments, quantum features may include quantum computing generally. For example, quantum features may include quantum optimization of power grids (e.g., a quantum computing system configured to optimize a power grid). Examples of quantum computing also include a quantum computing system configured for design optimization (e.g., optimizing the design of a battery) and/or a quantum computing system configured for real-time operating temperature optimization (e.g., optimizing the operating temperature of a battery). can do. In example embodiments, quantum computing for design or real-time operating temperature optimization may include automated design and system control.

In other example embodiments, quantum features may be used for quantum optimized battery disposal (e.g., quantum computing systems for optimizing battery disposal) and/or quantum optimized battery recycling or reuse (e.g., for optimizing battery recycling or reuse). It may include quantum battery optimization, such as quantum computing systems. In example embodiments, these quantum computing systems provide supply chain optimization for primary and intermediate battery materials that may involve integration of processes and more vertically integrated operations including battery production and juxtaposition of products. Surrounding concerns can be resolved. For example, these quantum computing systems could address the needs of battery manufacturing, which could include material use optimized with VCN modeling. These quantum computing systems can be utilized in battery manufacturing using 3D printing materials and processes. These quantum computing systems can also be used with data collection, management, and analysis that can integrate testing and tracking of battery “cells” and other subcomponents for VCN optimization. Quantum computing systems can be used for battery recycling or reuse, as well as with battery disposal, carbon footprint management, etc. (for example, the supply of lithium or cobalt needed for lithium-ion batteries can come from recycled materials). there is).

In example embodiments, quantum features may include quantum optimization of energy use mix (e.g., a quantum computing system for optimizing energy use mix). In other example embodiments, quantum features may include energy cost optimization across distributed commerce models (e.g., a quantum computing system configured to optimize energy cost across distributed commerce models). These quantum computing systems include several manufacturing locations and multiple types of routes and types of transportation (e.g., third party logistics (3PL), fourth party logistics (4PL), super grid logistics, logistics It can be used with VCNs that use markets). These quantum computing systems can optimize decisions about where/when to manufacture and how/when to transport according to the real-time and predicted cost of energy as input. This will be achieved by monitoring all entities across the VCN and analyzing different variables to ensure that customer demands are being met while keeping energy costs to a minimum (e.g., using AI, predictive analytics, and quantum). You can. The same can be applied to multi-tenancy facilities to provide customers with the opportunity to optimize their energy usage costs. Additionally, predicting and securing future energy demand can be based on expected customer demand.

In example embodiments, energy system 19000 includes technologies such as systems for slicing production, storage, and/or delivery (e.g., type/mix, time, and location tracking) of energy. can do.

In example embodiments, energy system 19000 may be utilized in a variety of use cases. In example embodiments, energy system or process 19000 may be used to process high-value items that may have local demand (e.g., growing high-margin foods, high-energy computational workloads, and/or high-temperature materials processes). It can be applied with a variety of use cases such as modular/small-scale energy supply systems and the juxtaposition of various production systems. In example embodiments, other use cases include moving energy storage devices (e.g., systems with a network of moving energy storage devices such as energy trucks), food-to-energy value chain networks, and micro-power plants. Fractional ownership (e.g., a system for tokenizing ownership stakes in micro-power plants so that their investment costs can be crowd-sourced among owners), solar panels and roads (e.g., integrated solar integration of roadways with optical panels), coordination systems of supply and demand intersections (e.g., geographic and land use permits, computing and/or data center resources, and both in time with energy availability), battery-based Includes a printed circuit board manufacturing plant (e.g., the energy may be generated from renewable energy sources), and/or an energy index (e.g., an energy index for determining the price of a product) can do. In some example embodiments, robotics technology and energy optimization technology may be used together (eg, to provide an energy-optimized platform for autonomous robotic operations).

In example embodiments, there may be various examples of fractional ownership of micro plants. Ownership shares of “micro-power stations” can be tokenized (e.g., through the use of “ownership” tokens), so that the investment costs of such renewable energy sources can be divided among owners. Can be crowd-sourced. Each micro-power plant may include renewable energy sources, such as solar panels, wind turbines, geothermal energy conversion, and/or energy storage (eg, large batteries). Additionally, property owners where micro-power plants may be located may also be awarded ownership tokens. Smart contracts can be created to release energy into the grid, whereby each micro-power plant (e.g., energy source/battery) can be “coin-operated.” Smart contracts can define prices for energy and contain conditional logic that can be configured to instruct energy stores/energy sources to release energy into the grid when the required funds are deposited into the account associated with the smart contract. can do. The smart contract can then distribute funds to the overall owners of the energy source, energy storage, and/or property owner based on the ownership tokens each owner owns (e.g., ownership of a micropower plant) Owners of ¼ of the tokens can receive ¼ of the distribution funds). The ownership tokens can then become tradable, assuming the micro-power plant is successful. This considers putting energy back into the “grid,” but other examples could apply to solutions for roadside vehicle chargers, industrial growth facilities, crypto mining facilities, etc. In some examples, builders of new facilities (e.g., industrial growth, power plants, crypto-mining facilities) can crowd-source electricity production without having to give up equity in their business or take out a loan. You can. In this scenario, excess energy could be sold to the grid, which would in turn increase the value of ownership tokens.

In example embodiments, there may be various examples of battery-based printed circuit board manufacturing plants. For example, a printed circuit board manufacturing plant can receive raw materials, produce millions of polychlorinated biphenyl (PCB) per day, have AI-enabled problem/defect identification and prediction, and transport PCBs to electric vehicles ( It can be shipped to ports, warehouses, distribution centers, etc. (for example, electric trucks and cargo vans). A manufacturing plant may have solar panels on top and near the plant building that can supply batteries stored within the plant. Wave energy harvesting devices could be installed on nearby ocean shores, where the energy harvested would be transmitted to plants and stored in batteries. A nearby field may have a wind farm stored thereon and the wind energy is transmitted to the battery. There may be several sets of batteries, each set of batteries capable of receiving power from one or more of solar panels, wave energy harvesting devices, and wind farms. One of the sets of batteries may be configured to power the manufacturing plant during operating hours. The second set of batteries, or a subset of the first set of batteries, may be configured to supply transformers configured to charge electric vehicles. Power from solar, wave and wind energy sources supplies high-power computing devices, such as electric vehicle (EV) charging transformers to manufacturing machines, and AI-enabled forecasting and plant management/control systems. By adjusting the phase, capacitance, etc., it can be adapted for storage in different types of batteries and for use in different applications.

Illustratively, there may be various examples using an energy index. For example, if this is a competition on energy efficiency between suppliers in a private network or supply chain, the energy index can be uniform or comparable across products (e.g. similar components that may go into the final product). can be added to determine the price of products). Goods that may have used more energy to manufacture (or manufacture and deliver) may be sold at a discount, while goods that may have used less energy to manufacture (or manufacture and deliver) may be sold at a premium. You can. Blockchain can be used to track batches rather than individual items. Verification (or certification) capabilities may be required. Energy sources can also be factors such as coal versus renewable potential. If multiple manufacturers of the end product are involved, anti-trust mechanisms may need to be in place. In some examples, these use cases may be coupled with energy production cost tracking technologies (e.g., to feed a blockchain).

Dual Process Artificial Neural Networks

Referring to FIG. 104 in conjunction with FIG. 165 , in embodiments, intelligent services include a dual process artificial neural network (DPANN) system 20000. The DPANN system 20000 includes an artificial neural network (ANN) with behavioral and behavioral processes (e.g., decision making) being the product of a training system and a retraining system. The training system is configured to perform automatic trained execution of ANN operations. The retraining system may include, for example, memory, one or more input data sets (including temporal information for elements within these data sets), one or more goals or objectives (e.g., periodically and/or those related to the context of use of the ANN) and Perform effective, analytical and intentional retraining of the ANN based on one or more relevant aspects of the ANN, such as those that may change dynamically based on context changes), and/or others. In cases involving memory-based retraining, the memory may include original/historical training data and improved training data. The DPANN system 20000 includes a dual process learning function (DPLF) configured to manage and perform the ongoing data retention process. DPLF (including memory management processes, if applicable) facilitates retraining and improvement of the behavior of the ANN. DPLF allows the ANN to process historical inputs, new inputs, and new outputs (parameters in the context of usage (latency parameters, accuracy parameters, consistency parameters, bandwidth utilization parameters, processing capacity utilization parameters, prioritization). predictions, classifications, and outputs configured for specific use cases, including those determined by parameters, which may include performance parameters such as energy usage parameters, and many others; Provides a framework for generating output such as recommendations, conclusions and/or other output.

In embodiments, DPANN system 20000 stores training data to allow constant retraining based on the results of the ANN's decisions, predictions, and/or other operations, as well as based on the outputs of the ANN. This allows analysis of training data. Management of entities stored in memory allows the construction and execution of new models, such as those that can be processed, executed, or otherwise performed by or under the management of a training system. The DPANN system (20000) uses instances of memory to verify actions (e.g., in a manner similar to the thinking of biological neural networks (including retrospective or self-reflective thinking about which actions were optimally performed under given circumstances). ), perform training of the ANN, which includes training by intentionally supplying the ANN with appropriate sets of memories (i.e., those that produce favorable results given the performance requirements for the ANN).

In embodiments, the DPLF may be or include a continued process retention of one or more training data sets and/or memories stored in memory over time. Thereby, DPLF enables an ANN to apply existing neural functions and understand its behavior within current, recent, and/or new scenarios, including in simulations of the ANN, during training processes, and in fully operational deployments. allows to derive sets of past events (including those that are intentionally altered and/or curated for distinct purposes), such as framing a DPLF can provide ANNs with a framework through which ANNs can analyze, evaluate and/or manage data, such as data related to the past, present and future. In this way, DPLF plays an important role in training and retraining ANN through training system and retraining system.

In embodiments, the DPLF is configured to perform dual process operation to manage existing training processes, and also to manage and/or perform new training processes, i.e., retraining processes. In embodiments, each instance of the ANN is configured to be trained via a training system and retrained via a retraining system. An ANN encodes training and/or retraining data sets, stores the data sets, and retrieves the data sets during both training with the training system and retraining with the retraining system. The DPANN system (20000) provides a data set (selectively including various subsets, supersets, combinations, permutations, elements, metadata, augmentations, etc. relative to the base data set used for training or retraining). The term data set in this context), storage activity, processing operation and/or output is based on its respective inputs, processing (e.g. its structure, type, models, operations, execution environment, resource utilization, etc. ) and/or outcomes (including outcome types, performance requirements (including context or dynamic requirements), etc.) to recognize whether a training system versus a retraining system has characteristics that inherently favor it. For example, the DPANN system (20000) suggests that poor performance of a training system on a classification task may be due to new problems for which the training of the ANN is not appropriate (e.g., type of data set, nature of input models, and/or feedback). , the amount of training data, the quality of tagging or labeling, the quality of supervision, etc.), and whether the processing operations of the ANN are not well suited (e.g., the type of neural network used, the type of models used, etc.) prone to vulnerabilities), can be addressed by engaging a retraining system to retrain the model to teach the model to learn to solve new classification problems (e.g., much better labeling of correctly classified items). With periodic or continuous evaluation of the performance of the ANN, the DPANN system can subsequently determine that only small improvements in the ANN are achieved over many iterations of retraining (e.g. by a retraining system). (or, if both occur, will be weighted more favorably) that the highly stable performance of the ANN indicates readiness for the training system to replace the retraining system (or, if both are involved, over longer time periods). , cycles of variable performance may emerge, as a series of new problems emerge, and thus the DPANN's retraining system can retrain the ANN as needed and/or augment the ANN by providing a second source of outputs. Serially involved (can be fused or combined with ANN outputs to provide a single result (with various weights across them)), or can involve comparison, selection, averaging, or context- or situation-specific application of individual outputs. (which can be served in parallel).

In embodiments, the ANN is configured to learn new functions with the collection of data following dual process training of the ANN through a training system and a retraining system. The DPANN system 20000 performs an analysis of the ANN through a training system and allows the ANN to acquire new internal functions (or internal functions are subtracted or modified, e.g., if existing functions do not contribute to favorable results). Carry out initial training. After initial training, the DPANN system 20000 performs retraining of the ANN through a retraining system. To perform retraining, the retraining system evaluates the ANN's memory and historical processing to construct targeted DPLF processes for retraining. DPLF processes may be specific to identified scenarios. The ANN process can run in parallel with the DPLF process. As an example, an ANN may function to operate a specific make and model of self-driving car after initial training by a training system. DPANN system 20000 may, for example, allow ANNs to detect different makes and models of automobiles (e.g., those with different cameras, accelerometers and other sensors, different physical characteristics, different performance requirements, etc.), or even bicycles or In order to be able to operate different types of vehicles, such as spacecraft, retraining of the functions of the ANN can be performed through a retraining system.

In embodiments, as the quality of the ANN's outputs and/or operations improves, and as long as the performance requirements for the ANN and the context of utilization remain fairly stable, performing a dual-process training process becomes less and less ( It can be a (decreaingly) tricky process. As such, the DPANN system 20000 requires fewer neurons in the ANN to perform the operations and/or processes of the ANN, performance monitoring can be less intensive (e.g., longer intervals between performance checks), and ), and/or may determine that retraining is no longer needed (at least for a period of time, such as until a long-term maintenance period is reached and/or until there are significant shifts in the context of use). As ANNs continue to improve existing features and/or add new features through a dual-process training process, they sometimes take on more “intellectually-demanding” (i.e., retraining-intensive) tasks simultaneously. It can be done. For example, by leveraging dual-process learning knowledge of the function or process being trained, ANNs can solve unrelated complex problems or make retraining decisions simultaneously. Retraining involves an agent (e.g., a human supervisor or an intelligent agent) instructing the ANN with a retraining goal (e.g., “master this new function”) and providing training tasks and feedback functions for retraining. May include supervision, such as providing a set of supervisory ratings (e.g., supervisory ratings). In embodiments, an ANN may be used to organize the supervision, training and retraining of other dual process trained ANNs, seed such training or retraining, etc.

In embodiments, one or more behaviors and operational processes (eg, decision making) of the ANN may be the products of training and retraining processes facilitated by a training system and a retraining system, respectively. The training system may be configured to perform automatic training of the ANN, such as by continuously adding additional instances of training data as collected by or from various data sources. A retraining system may be an efficient, analytical, purposeful retraining of an ANN, for example, based on memory (e.g., stored training data or refined training data) and/or optionally based on inference or other factors. It can be configured to perform. For example, in a deployment management context, a training system may be associated with standard responses by an ANN, while a retraining system may implement DPLF retraining and/or network adaptation of the ANN. In some cases, retraining of an ANN beyond the factory, or “out-of-the-box” training level may involve more than retraining by a retraining system. Successful tuning of an ANN by means of one or more network adaptations may depend on the operation of the one or more network adaptations of the training system.

In embodiments, a training system may facilitate fast operation and training of an ANN by applying existing neural functions of the ANN based on training of the ANN with previous data sets. Standard operational activities of an ANN, which may rely heavily on the training system, include, without limitation, one or more of the methods, processes, workflows, systems, etc. described throughout this disclosure and the documents included herein. May include: defined functions within networking (e.g., discovering available networks and connections, establishing connections in networks, provisioning network bandwidth between devices and systems, such as routing data within networks, directing traffic to available network paths, load balancing across networking resources, and many others); Recognition and classification (such as images, text, symbols, objects, video content, music and other audio content, speech content, and many others); spoken words; prediction of states and events (such as prediction of failure modes of machines or systems, prediction of events in workflows, predictions of behavior in shopping and other activities, and many others); Control (e.g., controlling autonomous or semi-autonomous systems, automated agents (e.g., automated call-center operations, chat bots, etc.) and others); and/or optimization and recommendations (e.g., for products, content, decisions, and many others). ANNs can also be suited to training data sets for scenarios that only require output. Standard operational activities will not require the ANN to actively analyze what the ANN is requesting beyond operating on well-defined data inputs to compute well-defined outputs for well-defined use cases. You can. The operations of the training system and/or retraining system may be based on one or more historical data training data sets and may use parameters of the historical data training data sets to calculate results based on new input values, such as an ANN or It can be performed with or without small changes to its input types. In embodiments, an instance of the training system may be configured to determine whether the ANN is capable of performing a task in a given situation, such as by recognizing whether the image or sound being classified by the ANN is of a type that has historically been classified with high accuracy (e.g., above a threshold). They can be trained to classify whether they can perform well or not.

In embodiments, network adaptation of an ANN by one or both the training system and the retraining system includes a number of defined network functions, knowledge, and intuition-like behavior of the ANN when subjected to new input values. can do. In these embodiments, the retraining system may perform retraining of the ANN by applying new input values to the DPLF system to adjust the functional response of the ANN. DPANN System (20000) includes, for example, and without limitation, functional neural networks being assigned activities and tasks that require ANNs to provide solutions to new problems, engage in network adaptation or other higher-order cognitive activities, and When a DPANN applies concepts outside the domain for which it was originally designed and supports different contexts of deployment (such as when use cases, performance requirements, available resources, or other factors have changed), the re-development of the ANN through network adaptation You may decide that training is needed. ANNs are susceptible to, for example, poor performance of the training system, high variability of input data sets relative to the historical data sets used to train the training system, new functional or performance requirements, and dynamic changes in use cases or context. By training the ANN to recognize , or other factors, a retraining system can be trained to recognize where it is needed. ANNs can apply inference to evaluate performance and provide feedback to the retraining system. ANN may have, for example, inputs (e.g., data sources), processes/functions (e.g., neural network types and structures), feedback, and outputs, or various permutations of these elements. and optionally by a combinatorial or recombinant process involving genetic programming arranged in combinations, the ANN may be trained and/or retrained (in simulations or live batches) ), with respect to each, are tested, e.g., in a series of rounds or evolutionary steps, promoting advantageous variations until a preferred ANN or preferred set of ANNs for a given scenario, use case, or set of requirements is identified. This may be achieved by retraining systems and subsequent training and/or Bayesian inference processes, causal or conditional inference processes, deductive inference processes, inductive inference processes, or other methods described in this disclosure or documents incorporated herein by reference ( may involve generating a set of input "ideas" (e.g., combinations of different conclusions about cause and effect in a diagnostic process) for processing by an explicit inference process (including combinations of the above). You can.

Referring to Figure 2xm, in embodiments, the DPLF may perform an encoding process of the DPLF to process the data sets into a stored form for future use, such as retraining of the ANN by a retraining system. The encoding process allows data sets to be taken, understood, and modified by the DPLF to better support storage in and use from memory. DPLF can apply current functional knowledge and/or reasoning to integrate new input values. Memory may include short-term memory (STM), long-term memory (LTM), or a combination thereof. Data sets may be stored in either or both STM and LTM. STM can be implemented by the application of specialized behaviors inside the ANN (e.g., recurrent neural networks, or long-short-term neural networks, which may be gated or ungated). LTM can be implemented by storing scenarios, associated data, and/or unprocessed data that can be applied to the discovery of new scenarios. The encoding process can, for example, combine visual encoding data (e.g. processed through a convolutional neural network), acoustic sensor encoded data (e.g. how something sounds), speech encoded data (e.g. (processed through a deep neural network (DNN), including phoneme recognition), for example, semantic encoding data of words to determine semantic meaning by using a Hidden Markov Model (HMM); and/ or processing and/or storing movement and/or tactile encoded data (e.g., motion for vibration/accelerometer sensor data, touch sensor data, location or geo-location data, etc.) Data sets may be used in any of these modes. Although they may enter the DPLF system through one, the format in which the data sets are stored may differ from their original form and pass-through neural processing engines to be encoded in a compressed and/or context-relevant format. For example, an unsupervised instance of an ANN can be used to learn historical data in a compressed format.

In embodiments, encoded data sets are maintained within the DPLF system. The encoded data set is first stored in short-term DPLF, or STM. For example, a sensor data set may be stored primarily in an STM and maintained in the STM through constant repetition. Data sets stored in STM are active and function as a kind of immediate response to new input. DPANN system 20000 may retrieve data sets from an STM in response to changes in data streams due to out of space in the STM, for example, when new data is imported, processed, and/or stored. It can be removed. For example, it is possible for a short-term DPLF to last only between 15 and 30 seconds. STM can store only a small amount of data, which is typically embedded within the ANN.

In embodiments, DPANN system 20000 may, for example, consume various indicators of attention and/or utilization of outputs from the ANN and transmit such indicators to the ANN in response (attention). ) is similar to the “awareness moment” in the brain where something is overlooked and the cognitive system says “aha!”), the training system, the DPANN system as a whole (20000), and/or similar can be measured. In embodiments, attention may be measured by the sheer amount of activity of one or both systems on the data stream. In embodiments, a system using output from an ANN may be implemented, e.g., by an operator instructing the ANN to attend to a particular activity (e.g., to respond to a diagnosed problem, among many other possibilities). Attention can be expressed explicitly. The DPANN system 20000 can be used to, for example, provide data with high intrinsic variability from historical patterns (e.g., in rates of change, departure from norm, etc.), data exhibiting high variability in historical performance ( data inputs to facilitate attention measures, for example, by prompting and/or calculating greater attention to data with similar characteristics to the data sets involved in situations in which the ANN performed poorly in training. It can be managed.

In embodiments, DPANN system 20000 may retain encoded data sets within a DPLF system pursuant to and/or as part of one or more storage processes. The DPLF system stores the encoded data sets in the STM and stores the encoded data sets in the LTM as needed after it is determined that they are no longer needed and/or have low priority for the current operation of the ANN, training process, retraining process, etc. You can save it. LTM can be implemented by storing scenarios, and DPANN system 20000 can apply associated and/or raw data to discovery of new scenarios. For example, data from a particular processed data stream, such as a semantically encoded data set, may be primarily stored in the LTM. LTM can also store image (and sensor) data sets in encoded form, among many other examples.

In embodiments, the LTM may have a relatively high storage capacity, and data sets stored within the LTM may effectively be stored indefinitely in some scenarios. DPANN system 20000 can, for example, pass LTM data through a series of memory structures with increasingly longer retrieval periods or increasingly higher threshold requirements to trigger utilization (as the biological brain processes challenging problems). (akin to "thinking very hard" to find precedents that do this), it can be configured to remove data sets from the LTM, thereby removing old data sets (with more time/effort) when the situation justifies more comprehensive memory utilization. Provides increased salience of more recent or more frequently used memories while maintaining the ability to retrieve memories. As such, the DPANN system 20000 can store older memories (as measured by time of origin and/or latest time of use) on a separate and/or slower system. Data stored in LTM on a timeline by imposing penalties by imposing artificial delays on their retrieval, and/or by imposing threshold requirements (e.g., indicators of high demand for improved results) before use. Sets can be arranged. Additionally or alternatively, LTMs may be clustered according to other classification protocols, such as by topic. For example, all memories that are temporally proximate to a periodically recognized person may be clustered together for search and/or all memories associated with a scenario may be clustered together for search.

In embodiments, the DPANN system 20000 modularizes and links LTM data sets, e.g., in catalogs, hierarchies, clusters, knowledge graphs (with directed/acyclic or conditional logic), etc. It can facilitate search. For example, all memory modules with instances containing people, topics, items, processes, concatenations of n-tuples of such (e.g., all memory modules containing selected pairs of entities), etc. field. The DPANN system 20000 may be used in one or more domains, such as training a model to predict robot or human agent behavior by using memories associated with a particular set of robots or human agents, and/or similar robots or human agents. You can select sub-graphs of the knowledge graph for DPLF to implement with specific and/or task-specific uses. The DPLF system may cache frequently used modules due to different speeds and/or availability. High-value modules (eg, those with high quality results, performance characteristics, etc.) can be used for other functions, such as selection/training of STM keep/forget processes.

In embodiments, DPANN system 20000 may modularize and link LTM data sets, such as in the various ways mentioned above, to facilitate retrieval of related memories. For example, an instance involving a person, topic, item, process, a concatenation of n-tuples of such (e.g., all memory modules containing a selected pair of entities), or all memories associated with a scenario, etc. Memory modules having them can be connected and searched. DPANN system 20000 is a domain-specific and /Or one may select subsets of scenarios for the DPLF (e.g., subgraphs of the knowledge graph) for task-specific use. Frequently used modules or scenarios may be stored in the cache for speed/availability or other performance characteristics. High-value modules or scenarios (scenarios with high-quality results) can be used for other functions, especially selection/training of STM archiving/forgetting processes.

In embodiments, DPANN system 20000 may perform LTM planning, for example, to find a procedural course of action for the declaratively described system to reach its goals while optimizing overall performance measures. . The DPANN system 20000 can, for example, solve problems where the problem can be described in a declarative manner, the DPANN system 20000 has domain knowledge that should not be ignored, and the structure of the problem makes it difficult for pure learning techniques. and/or perform LTM planning when the ANN needs to be trained and/or retrained to be able to account for specific courses of action taken by the DPANN system 20000. In embodiments, the DPANN system 20000 can be applied to a plan recognition problem, i.e., the inverse of a planning problem: instead of a goal state, a set of possible goals is given, and the goal in plan recognition is which goals are being achieved. and finding out how.

In embodiments, DPANN system 20000 may facilitate LTM scenario planning by a user to develop a long-term plan. For example, LTM scenario planning for risk management use cases is not typically considered in day-to-day operations, such as those outside the bell curve or normal distribution, among many others, but in fact information or market Extreme or unusual, but likely, risks and opportunities that occur with greater frequency than expected in “long tail” or “fat tail” situations, such as those involving pricing processes. Additional emphasis may be placed on discernment. LTM scenario planning may involve analyzing relationships between influences (such as social, technological, economic, environmental, and/or political trends) to describe current conditions and/or potential future states. This may include providing scenarios for

In embodiments, DPANN system 20000 may facilitate LTM scenario planning to predict and anticipate possible alternative futures with the ability to respond to predicted states. LTM plans can be derived from expert domain knowledge or can be projected from current scenarios, because many scenarios (such as those involving the results of combinatorial processes that result in new entities or behaviors) have not yet occurred and Therefore, it cannot be projected by probabilistic means that depend entirely on historical distributions. The DPANN system 20000 can prepare applications for LTM to generate many different scenarios, exploring a variety of possible futures for DPLM, both expected and unexpected futures. This can be particularly facilitated or augmented by genetic programming and inference techniques as mentioned above.

In embodiments, DPANN system 20000 may implement LTM scenario planning to facilitate transforming risk management into a plan awareness problem and apply DPLF to generate possible solutions. LTM scenario derivation solves several problems that arise in planning forecasts. LTM scenario derivation occurs, for example, when the models used for prediction have inconsistent, missing, or unreliable observations; When it is possible to generate not just one, but many future plans; and/or may be applicable when LTM domain knowledge can be captured and encoded to improve predictions (e.g., when domain experts tend to outperform available computational models). LTM scenarios can focus on applying LTM scenario planning for risk management. LTM scenario planning can provide situational awareness of relevant risk drivers by detecting new storylines. LTM scenario planning can also create future scenarios that allow DPLM or operators to account for and plan for future contingencies and opportunities.

In embodiments, DPANN system 20000 may be configured to perform a search process through the DPLF to access the ANN's stored data sets. The search process can determine how well the ANN performs on assignments designed to test recall. For example, an ANN can be trained to perform controlled vehicle parking operations, whereby the autonomous vehicle returns to a designated point or exit by associating previous visits through retrieval of data stored in the LTM. Data sets stored in STM and LTM can be retrieved by different processes. Data sets stored in an STM may be retrieved in response to specific input and/or by the order in which the data sets are stored, for example, by a sequential numbered list. Data sets stored in the LTM may be searched through association and/or matching of events to historical activities, for example, through complex associations and indexing of large data sets.

In embodiments, DPANN system 20000 may implement scenario monitoring as at least part of the search process. A scenario can provide context for context decision-making processes. In embodiments, the scenarios may involve explicit inference (e.g., cause-and-effect inference, Bayesian, causal, conditional logic, etc., or combinations thereof), the output of which LTM-stored data Declares whether the timeline of events being evaluated and potentially other timelines involving events that follow similar cause-and-effect patterns are being searched. For example, diagnosis of a failure of a machine or workflow may be performed using historical sensor data as well as the various failure modes (and/or diagnosis of problem states or conditions, recognition of events or behavior, failure modes ( You can retrieve LTM data about similar processes (for example, financial breakdowns, breach of contract, etc.), or many other things.

CONCLUSION

Background information is presented merely for context and is not necessarily well-known, routine, or conventional. Additionally, the background description is not an admission of what may or may not qualify as prior art. In fact, some or all of the background description may be the work of the named inventor, who is not otherwise known in the art.

Physical (e.g., spatial and/or electrical) and functional relationships between elements (e.g., between modules, circuit elements, semiconductor layers, etc.) are described using various terms. Unless explicitly described as being “direct,” when a relationship between a first element and a second element is described, that relationship (i) is defined as such: (i) no other intervening elements exist between the first element and the second element; It includes both direct relationships and (ii) indirect relationships where one or more intervening elements exist between the first element and the second element. Exemplary relationship terms include “attached,” “sending,” “receiving,” “connected,” “involved,” “coupled,” “adjacent,” “next to,” “above,” “above,” and “below.” Includes “on,” “adjacent to,” and “placed on.”

The detailed description includes specific examples for illustrative purposes only and does not limit the disclosure or its applicability. The examples are not intended to be an exhaustive list, but rather are simply intended to demonstrate that the full scope of the presently presented and contemplated future claims belongs to the inventor. Modifications, combinations and equivalents of the examples are within the scope of this disclosure. No language in the specification should be construed as indicating that any non-claimed element is essential or critical to the practice of the disclosure.

The term “exemplary” simply means “example” and does not refer to the best or preferred example. The term "set" does not necessarily exclude empty sets, that is, in some situations a "set" may have zero elements. The term "non-empty set" may be used to indicate the exclusion of an empty set, i.e., a non-empty set must have at least one element. The term "subset" does not necessarily require a true subset. That is, a “subset” of the first set may span the same space as the first set. Additionally, the term "subset" does not necessarily exclude empty sets in some situations, and a "subset" may have zero elements.

The phrase "at least one of A, B, and C" shall be interpreted to mean logic (A or B or C) using the non-exclusive logical OR, and "at least one A, at least one B, and at least It should not be interpreted to mean "one C". The use of singular terms and similar referents in the context of describing the present disclosure and claims includes both the singular and the plural, unless explicitly or contradictory by context. Unless otherwise specified, the terms “comprising,” “having,” “with,” “including,” and “containing,” and their Variants are open-ended terms meaning “including, but not limited to.”

Each publication referenced in this disclosure, including foreign and domestic patent applications and patents, is hereby incorporated by reference in its entirety.

Although each of the embodiments has been described above as having specific features, any one or more of such features described in connection with any embodiment of the present disclosure may be used in any of the other embodiments, even if the combination thereof is not explicitly described. It may be implemented in the features of and/or combined with it. That is, the described embodiments are not mutually exclusive, and permutations of many embodiments remain within the scope of the present disclosure.

One or more elements (e.g., steps, instructions, actions, or operations) within a method may be executed in a different order (and/or simultaneously) without changing the principles of the present disclosure. Unless technically infeasible, elements described as serial may be implemented partially or fully in parallel. Similarly, unless technically infeasible, elements described as parallel may be implemented partially or fully serially.

Although this disclosure describes structures that correspond to claimed elements, such elements do not necessarily invoke a functional interpretation of the claims unless the "means for" description is explicitly used. Unless otherwise indicated, recitations of ranges of values are intended to serve only as a shorthand way of referring individually to each individual value falling within the range, and each individual value is incorporated into the specification as if it were individually recited. .

The drawings divide elements of the disclosure into different functional blocks or action blocks, but this division is for illustrative purposes only. According to the principles of the present disclosure, functionality can be combined in different ways such that some or all functionality from multiple individually shown blocks can be implemented in a single functional block; Similarly, functionality displayed in a single block may be separated into multiple blocks. Unless explicitly stated to be mutually exclusive, features shown in different drawings may be combined consistent with the principles of the disclosure.

In the drawings, reference numbers may be reused to identify identical elements or may simply identify elements that implement similar functionality. The numbering or other labeling of instructions or method steps is intended for convenient reference and not to indicate a fixed order. In the drawings, the direction of arrows, as indicated by arrow heads, generally shows the flow of information (such as data or instructions) relevant to the example. For example, when element A and element B exchange various information, but information transmitted from element A to element B is relevant for the example, an arrow may point from element A to element B. This one-way arrow does not imply that no other information is transmitted from element B to element A. As just one example, for information sent from element A to element B, element B may send a request and/or acknowledgment to element A.

Special purpose systems include hardware and/or software and may be described in terms of devices, methods, or computer-readable media. In various embodiments, functionality may be allocated differently between software and hardware. For example, some functionality may be implemented by hardware in one embodiment and by software in another embodiment. Additionally, software can be encoded by hardware structures, and hardware can be defined by software, as in software-defined networking or software-defined radio.

In this application, including the claims, the term module refers to a special purpose system. A module may be implemented by one or more special-purpose systems. One or more special-purpose systems may also implement some or all of the other modules. In this application, including the claims, the term module may be replaced by the term controller or circuit. In this application, including the claims, the term platform refers to one or more modules that provide a set of functionality. In this application, including the claims, the term system may be used interchangeably with the term module or special purpose system.

Special purpose systems can be directed or controlled by an operator. A special purpose system may be hosted by one or more of assets owned by the operator, assets leased by the operator, and third party assets. The asset may be referred to as a private, community, or hybrid cloud computing network or cloud computing environment. For example, special-purpose systems may be partially or fully hosted by third parties providing software as a service (SaaS), platform as a service (PaaS), and/or infrastructure as a service (IaaS). . Special-purpose systems can be implemented using agile DevOps (agile development and operations) principles. In embodiments, some or all of a special purpose system may be implemented in a multi-environment architecture. For example, multiple environments may include one or more production environments, one or more integration environments, one or more development environments, etc.

A special purpose system may be implemented partially or fully using or by a mobile device. Examples of mobile devices include navigation devices, cell phones, smartphones, mobile phones, mobile personal digital assistants, palmtops, netbooks, pagers, e-book readers, tablets, music players, etc. A special purpose system may be implemented partially or fully using or by a network device. Examples of network devices include switches, routers, firewalls, gateways, hubs, base stations, access points, repeaters, head-ends, user equipment, cell sites, antennas, towers, etc.

Special purpose systems can be partially or fully implemented using computers of various form factors and different characteristics. For example, a computer may be characterized as a personal computer, server, etc. The computer may be portable, as in the case of a laptop, netbook, etc. A computer may or may not have any output devices, such as a monitor, line printer, liquid crystal display (LCD), light emitting diode (LED), etc. A computer may or may not have any input devices such as a keyboard, mouse, touchpad, trackpad, computer vision system, barcode scanner, button array, etc. The computer may run a general-purpose operating system, such as a variant of the WINDOWS operating system from Microsoft Corporation, the MACOS operating system from Apple, Inc., or the LINUX operating system. Examples of servers include file servers, print servers, domain servers, Internet servers, intranet servers, cloud servers, infrastructure-as-a-service servers, platform-as-a-service servers, web servers, secondary servers, host servers, distributed servers, failover servers, and Includes backup server.

The term hardware includes components such as processing hardware, storage hardware, networking hardware, and other general-purpose and special-purpose components. Note that these are not mutually exclusive categories. For example, processing hardware may integrate storage hardware, and vice versa.

Examples of components include integrated circuits (ICs), application-specific integrated circuits (ASICs), digital circuit elements, analog circuit elements, combinational logic circuits, gate arrays such as field programmable gate arrays (FPGAs), digital signal processors (DSPs), and complex Programmable logic devices (CPLD), etc.

Multiple components of hardware may be integrated, for example, on a single die, within a single package, or on a single printed circuit board or logic board. For example, multiple components of hardware may be implemented as a system-on-chip. A component, or set of integrated components, may be referred to as a chip, chipset, chiplet, or chip stack. Examples of system-on-chip include radio frequency (RF) system-on-chip, artificial intelligence (AI) system-on-chip, video processing system-on-chip, organs-on-chip, and quantum algorithm system-on-chip. Including chips, etc.

Hardware may integrate and/or receive signals from sensors. Sensors measure temperature, pressure, wear, light, humidity, strain, expansion, contraction, deflection, bending, stress, strain, load-bearing, contraction, power, energy, mass, position, temperature, humidity, pressure, viscosity, and liquid flow. , may allow for the observation and measurement of conditions including chemical/gas presence, sound, and air quality. The sensor may include image and/or video capture in visible and/or invisible (e.g., thermal) wavelengths, such as a charge-coupled device (CCD) or complementary metal-oxide semiconductor (CMOS) sensor.

Examples of processing hardware include central processing units (CPUs), graphics processing units (GPUs), approximate computing processors, quantum computing processors, parallel computing processors, neural network processors, signal processors, digital processors, data processors, embedded processors, microprocessors, and Includes a coprocessor. A coprocessor may provide additional processing capabilities and/or optimizations, such as for speed or power consumption. Examples of coprocessors include math coprocessors, graphics coprocessors, communications coprocessors, video coprocessors, and artificial intelligence (AI) coprocessors.

A processor may enable execution of multiple threads. These multiple threads may correspond to different programs. In various embodiments, a single program may be implemented as multiple threads by a programmer or may be decomposed into multiple threads by processing hardware. Threads can run concurrently to improve processor performance and facilitate concurrent operation of applications. A processor may be implemented as a packaged semiconductor die. A die contains one or more processing cores and may include additional functional blocks such as a cache. In various embodiments, a processor may be implemented by multiple dies that may be combined into a single package or packaged individually.

Networking hardware may include one or more interface circuits. In some examples, the interface circuit(s) may implement a wired or wireless interface that connects directly or indirectly to one or more networks. Examples of networks include cellular networks, local area networks (LANs), wireless personal area networks (WPANs), metropolitan area networks (MANs), and/or wide area networks (WANs). The network may include one or more of point-to-point and mesh technologies. Data transmitted or received by a networking component may traverse the same or different networks. Networks can be connected to each other over WANs or point-to-point leased lines using technologies such as Multiprotocol Label Switching (MPLS) and virtual private networks (VPN).

Examples of cellular networks include GSM, GPRS, 3G, 4G, 5G, LTE, and EVDO. Cellular networks may be implemented using frequency division multiple access (FDMA) networks or code division multiple access (CDMA) networks. Examples of LANs are Institute of Electrical and Electronics Engineers (IEEE) Standard 802.11-2020 (also known as the WIFI wireless networking standard) and IEEE Standard 802.3-2018 (also known as the ETHERNET wired networking standard). Examples of WPAN include IEEE standard 802.15.4, which includes the ZIGBEE standard from the ZigBee Alliance. Other examples of WPANs include the Bluetooth wireless networking standard, including Core Specification versions 3.0, 4.0, 4.1, 4.2, 5.0, and 5.1 from the Bluetooth Special Interest Group (SIG). WAN may also be referred to as a distributed communications system (DCS). One example of a WAN is the Internet.

Storage hardware is or includes computer-readable media. The term computer-readable media, as used in this disclosure, includes both volatile and non-volatile storage, such as dynamic random access memory (DRAM). The term computer-readable medium excludes only transient electrical or electromagnetic signals propagating through the medium (e.g., on a carrier wave). Accordingly, computer-readable media in this disclosure may be considered non-transitory and also tangible.

Examples of storage implemented by storage hardware include databases (e.g., relational databases or NoSQL databases), data stores, data lakes, column stores, and data warehouses. Examples of storage hardware include non-volatile memory devices, volatile memory devices, magnetic storage media, storage area network (SAN), network-attached storage (NAS), optical storage media, printing media (such as barcodes and magnetic ink), and ( Includes paper media (such as punch cards and paper tape). Storage hardware may include cache memory that may be co-located or integrated with processing hardware. Storage hardware may have read-only, write-once, or read/write characteristics. Storage hardware may be random access or sequential access. Storage hardware may be location-addressable, file-addressable, and/or content-addressable.

Examples of non-volatile memory devices include flash memory (including NAND and NOR technologies), solid state drives (SSD), erasable programmable read-only memory devices such as electrically erasable programmable read-only memory (EEPROM) devices, and ROM. Includes (mask read-only memory device). Examples of volatile memory devices include random access memory (RAM) and processor registers such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), synchronous graphics RAM (SGRAM), and video RAM (VRAM). do. Examples of magnetic storage media include analog magnetic tape, digital magnetic tape, and rotating hard disk drives (HDD). Examples of optical storage media include CDs (eg, CD-R, CD-RW, or CD-ROM), DVDs, Blu-ray Discs, and Ultra HD Blu-ray Discs.

Examples of storage implemented by storage hardware include distributed ledgers, such as permissioned or permissionless blockchains. As in blockchain, the entities recording transactions can reach consensus using algorithms such as proof-of-stake, proof-of-work, and proof-of-storage. Elements of the present disclosure may be represented by or encoded as a non-fungible token (NFT). Ownership rights associated with non-fungible tokens may be recorded in or referenced by a distributed ledger. Transactions initiated by or related to this disclosure may use one or both fiat currency and cryptocurrency, examples of which include bitcoin and ether. Some or all features of the hardware may be defined using languages for hardware description, such as IEEE Std 1364-2005 (commonly referred to as "Verilog") and IEEE Std 1076-2008 (commonly referred to as "VHDL"). You can. A hardware description language can be used to manufacture and/or program hardware.

Special-purpose systems can be distributed across multiple different software and hardware entities. Communication within and between special purpose systems can be accomplished using networking hardware. The distribution may vary across embodiments and may change over time. For example, the distribution may vary based on demand, and additional hardware and/or software entities are invoked to handle higher demands. In various embodiments, a load balancer may direct requests to one of multiple instantiations of a special-purpose system. Hardware and/or software entities may be physically distinct and/or may share some hardware and/or software, such as in a virtualized environment. Multiple hardware entities may be referred to as server racks, server farms, data centers, etc.

Software includes machine readable and/or executable instructions. Instructions can be logically grouped into programs, code, methods, steps, actions, routines, functions, libraries, objects, classes, etc. Software may be stored by storage hardware or encoded on other hardware. The software consists of (i) descriptive text to be parsed, such as hypertext markup language (HTML), extensible markup language (XML), and JavaScript Object Notation (JSON), (ii) assembly code, and (iii) source code generated by a compiler. (iv) source code for execution by an interpreter, (v) byte code, (vi) source code for compilation and execution by a just-in-time compiler, etc. By way of example only, source code may be written using syntax from languages including C, C++, JavaScript, Java, Python, R, etc.

Software also includes data. However, data and instructions are not mutually exclusive categories. In various embodiments, instructions may be used as data in one or more operations. As another example, instructions may be derived from data. The functional blocks and flow diagram elements in this disclosure serve as software specifications, which can be converted into software by the routine work of a skilled technician or programmer. Software includes firmware, processor microcode, operating system (OS), basic input/output system (BIOS), application programming interfaces (APIs), libraries such as dynamic-link libraries (DLLs), device drivers, hypervisors, user applications, May include and/or rely on background services, background applications, etc. Software includes native applications and web applications. For example, a web application can be served to a device through a browser using HTML5 (hypertext markup language 5th revision).

The Software may contain artificial intelligence systems, which may include machine learning or other computational intelligence. For example, artificial intelligence may include one or more models used in one or more problem domains. When presented with many data features, identification of a subset of features related to the problem domain can improve prediction accuracy, reduce storage space, and increase processing speed. This identification may be referred to as feature engineering. Feature engineering can be performed by the user or guided only by the user. In various implementations, a machine learning system may computationally identify relevant features, such as performing singular value decomposition on the contributions of different features to the output.

Examples of models include recurrent neural networks (RNN) such as long short term memory (LSTM), transformers, decision trees, support-vector machines, genetic algorithms, Bayesian networks, and deep learning models such as regression analysis. Examples of systems based on the transformer model include bidirectional encoder representations from transformers (BERT) and generative pre-trained transformers (GPT). Training a machine learning model may include supervised learning (e.g., based on labeled input data), unsupervised learning, and reinforcement learning. In various embodiments, machine learning models may be pre-trained by the operator or by a third party. Problem domains include almost any situation in which structured data can be collected, and include natural language processing (NLP), computer vision (CV), classification, image recognition, etc.

Some or all of the software may run in a virtual environment rather than directly on hardware. The virtual environment may include a hypervisor, emulator, sandbox, container engine, etc. Software can be built as a virtual machine, container, etc. Virtualized resources may be controlled using, for example, the DOCKER container platform, the Pivot Cloud Foundry (PCF) platform, etc.

In the client-server model, some portions of the software run on a first piece of hardware that is functionally identified as a server, while other pieces of software run on a second piece of hardware that is functionally identified as a client. The identities of the client and server are not fixed: for some functions the first hardware may act as a server, while for other functions the first hardware may act as a client. In different embodiments and in different scenarios, functionality may be shifted between client and server. In one dynamic example, some functionality normally performed by the second hardware is shifted to the first hardware when the second hardware has less capability. In various embodiments, the term “local” may be used in place of “client” and the term “remote” may be used in place of “server.”

Some or all of the software can be logically partitioned into microservices. Each microservice provides a reduced subset of functionality. In various embodiments, each microservice can be scaled independently depending on the load, either by devoting more resources to the microservice or by instantiating more instances of the microservice. In various embodiments, the functionality provided by one or more microservices may be combined with each other and/or with other software that does not conform to the microservice model.

Some or all of the software can be logically arranged into layers. In a layered architecture, the second layer may be logically placed between the first and third layers. The first and third layers will then generally interact with the second layer without interacting with each other. In various embodiments, this is not strictly enforced - that is, some direct communication may occur between the first and third layers.

Claims (88)

In a digital product network system,
A set of digital products each having a product processor, product memory, and product network interface; and
A product network control tower having a control tower processor, a control tower memory, and a control tower network interface,
The product processor and the control tower processor,
generate product level data at the product processor;
transmit the product level data from the product network interface;
receive the product level data at the control tower network interface;
encode the product level data into a product level data structure configured to convey parameters indicated by the product level data across the set of digital products; and
A digital product network system, collectively comprising non-transitory instructions that program the digital product network system to write the product level data structure to at least one of the product memory and the control tower memory. 2. The digital product network system of claim 1, wherein the product network control tower is at least one of a remotely located server or at least one control product of the set of digital products. 2. The method of claim 1, wherein the product processor and the control tower processor are based on a shared communication system configured to facilitate communication of the product level data from the set of digital products with the product network control tower and between the digital products. A digital product network system that is further programmed to communicate with each other. 2. The digital product network system of claim 1, wherein the set of digital products and the product network control tower have a set of microservices and a microservices architecture. 2. The system of claim 1, further comprising a display associated with at least one of the product network control tower or the set of digital products, the digital product network system comprising:
create a graphical user interface having at least one user interface display;
generate parameters of at least one digitally enabled product of the set of digital products in the at least one user interface display; and
and further programmed to generate a proximity display of proximal digital products of the set of digital products in the at least one user interface display. 6. The method of claim 5, wherein generating the proximity display includes generating a proximity display of geographically proximate products, and the digital product network system is configured to display product types, product capabilities, and ), or a digital product network system further programmed to filter the proximal products by at least one of the product brands. 6. The method of claim 5, wherein generating the proximity display comprises proximity products that are in close proximity to one of the set of digital products by product type proximity, product capability proximity, or product brand proximity. A digital product network system, including generating a display. 2. The digital product network system of claim 1, wherein the digital product network system is further programmed to define a data integration system. 2. The digital product network system of claim 1, wherein the digital product network system is further programmed to provide edge computation and edge intelligence configured for edge distributed decision making among the set of digital products. 2. The digital product network system of claim 1, wherein the digital product network system is further programmed to provide edge intelligence and edge computation configured for edge network bandwidth management between or outside the set of digital products. 2. The digital product network system of claim 1, wherein the digital product network system is further programmed to have a distributed ledger system. 12. The digital product network system of claim 11, wherein the distributed ledger system is a blockchain ledger. 2. The digital product network system of claim 1, wherein the digital product network system is further programmed to have a quality management system with a system for capturing product complaints in the set of digital products. The method of claim 1, wherein the digital product network system:
identify the status of the set of digital products;
encode the state into one of the parameters of the product level data structure; and
A digital product network system, further programmed to at least one of: track or monitor the condition across the set of digital products. 2. The digital product network system of claim 1, wherein the digital product network system is further programmed to have a smart contract system to enable creation of smart contracts based on the product level data structure. 16. The digital product network system of claim 15, wherein the smart contract terms and conditions depend on the proximity of a plurality of digital products of the set of digital products. A digital product network system, further programmed to configure said smart contracts based on location-sensitive configuration. 2. The digital product network system of claim 1, wherein the digital product network system is further programmed to have a robotic process automation (RPA) system configured to gamify interaction based on which digital products are within the set of digital products. becoming a digital product network system. 2. The digital product network of claim 1, wherein the digital product network system has a robotic process automation (RPA) system and is further programmed to generate RPA processes based on use of a plurality of digital products of the set of digital products. system. In a digital product network system,
A set of digital products each having a product memory, a product network interface, and a product processor programmed with product instructions;
A product network control tower having a control tower memory, a control tower network interface, and a control tower processor programmed with control tower instructions; and
By at least one of the product instructions or the control tower instructions,
A digital product network system, comprising a digital twin system defined at least in part to encode a set of digital twins representing the set of digital products. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to encode hierarchical digital twins. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to encode a set of composite digital twins, each consisting of a set of discrete digital twins of the set of digital products. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to encode a set of digital product digital twins representing a plurality of digital products of the set of digital products. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to model traffic of moving elements within the set of digital products. 20. The method of claim 19, wherein the digital twin system comprises a set of digital twins where a user can replay data about a situation in the digital twin system and observe visual representations of events related to the situation. A digital product network system, further defined as having a playback interface. The method of claim 19, wherein the digital twin system,
create an adaptive user interface; and
further define to adapt at least one of the available data, features, or visual representations to the adaptive user interface based on at least one of the user's association or proximity to digital products of the set of digital products. , digital product network system. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to manage interactions between multiple digital product digital twins of the set of digital twins. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to create and update a self-expanding digital twin associated with the set of digital products. The method of claim 19, wherein the digital twin system,
Aggregating performance data from a plurality of digital twins among a set of digital twins for a common asset type represented in the plurality of digital twins; and
A digital product network system further defined to associate said aggregated performance data to a performance data set for retrieval. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to match owners of identical or similar products in the marketplace for digital twin data. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to lock the set of digital twins upon detection of a security threat in a digital product of the set of digital products. 20. The digital product network system of claim 19, wherein the digital twin system is further defined as having an in-twin marketplace. 32. The digital product network system of claim 31, wherein the in-twin marketplace is configured to provide data. 20. The digital product network system of claim 19, wherein the in-twin marketplace is configured to provide services. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to provide components. 20. The digital product network system of claim 19, wherein the digital twin system is further defined to include application program interfaces (APIs) between the set of digital twins and marketplaces associated with the set of digital products. 20. The method of claim 19, wherein the digital twin system is a twin store market system for providing at least one of access to or rights to at least one of the set of digital twins or data associated with the set of digital twins. ), which is further defined as having a digital product network system. In the autonomous futures contract orchestration platform,
A set of one or more processors programmed with a set of non-transitory computer-readable instructions, the non-transitory computer-readable instructions comprising:
receive, from a data source, an indication associated with a product associated with at least one entity that is buying or selling the product;
predict at least one baseline cost of buying or selling the product at a future time based on the indication;
retrieve, at a current time, the futures cost of a futures contract for an obligation to either buy or sell the product for at least one of delivery or performance of the product at the future time;
execute a smart contract for the futures contract based on the base cost and the futures cost; and
An autonomous futures contract orchestration platform that collectively executes orchestrating at least one of delivery or fulfillment of said product at said future point in time. According to clause 37,
further comprising a risk data structure representing the amount of risk the entity is willing to take for the base cost and the futures cost;
The computer readable instructions are:
An autonomous futures contract orchestration platform that collectively performs executing the smart contract based on the risk data structure for at least one of risk management or mitigation. 38. The autonomous futures contract orchestration platform of claim 37, further comprising a robotic process automation system for demand-side planning to orchestrate the smart contract. 38. The autonomous futures contract orchestration platform of claim 37, further comprising a robotic agent configured to derisk the futures contract and the smart contract. 38. The autonomous futures contract orchestration platform of claim 37, further comprising a system for performing circular economy optimization based on the futures price of the commodity. 38. The method of claim 37, wherein the computer readable instructions are:
initialize a robotic process automation system trained to execute the smart contract; and
An autonomous futures contract orchestration platform that collectively performs execution of the smart contracts using the robotic process automation system. 38. The autonomous futures contract orchestration platform of claim 37, wherein the indication is at least one of the occurrence of an event, the physical state of an item, or a potential increase in demand. In the autonomous futures contract orchestration platform,
A set of one or more processors programmed with a set of non-transitory computer-readable instructions, the non-transitory computer-readable instructions comprising:
Retrieving, at a current time, the futures cost of a futures contract for an obligation to buy or sell a product for at least one of delivery or performance of the product to an entity at a future time;
predict a basic cost for at least one of the entity to buy or sell the product at the future time based on an indication associated with the product associated with the entity;
execute a smart contract for the futures contract based on the base cost and the futures cost; and
An autonomous futures contract orchestration platform that collectively performs orchestrating at least one of the delivery or fulfillment of the product to the entity at the future point in time. A method for transmitting a predictive model of a data stream from a first device to a second device, the method comprising:
Receiving, by a first device, a plurality of data values of a data stream, the plurality of data values received comprising sensor data collected from one or more sensor devices;
generating, by the first device, a predictive model for predicting future data values of the data stream based on the received plurality of data values, wherein generating the predictive model comprises determining a plurality of model parameters; Contains steps -;
transmitting, by the first device, the plurality of model parameters to the second device;
receiving, by the second device, the plurality of model parameters;
parameterizing, by the second device, the prediction model using the plurality of model parameters; and
Predicting, by the second device, future data values of the data stream using the parameterized prediction model. 46. The method of claim 45, wherein the plurality of model parameters comprise vectors. 47. The method of claim 46, wherein the vector is a motion vector associated with a robot. 48. The method of claim 47, wherein the future data values of the data stream include one or more future predicted positions of the robot. 46. The method of claim 45, wherein the predictive model predicts stock levels of items, and the method further comprises:
based on the future data values, detecting an upcoming supply shortage of an item; and
The method further comprising taking steps to prevent shortage of the item. 46. The method of claim 45, wherein the prediction model is a behavior analysis model and the future data values are indicative of predicted behavior of the entity. 46. The method of claim 45, wherein the prediction model is an augmentation model and the future data values correspond to an inoperative sensor. 46. The method of claim 45, wherein the prediction model is a classification model and the future data values represent a predicted future state of a system including the one or more sensor devices. 46. The method of claim 45, wherein the one or more sensor devices are RFID sensors associated with cargo, and wherein the future data values indicate a future location of the cargo. 46. The method of claim 45, wherein the one or more sensor devices are security cameras and the data stream includes motion vectors extracted from video data captured by the security cameras. 46. The method of claim 45, wherein the one or more sensor devices are vibration sensors that measure vibrations produced by machines, and wherein the future data values indicate a potential need for maintenance of the machines. According to clause 45,
receiving, by the first device, additional data values of the data stream;
improving, by the first device, a predictive model using the additional data values, wherein improving the predictive model adjusts a plurality of model parameters; and
The method further comprising transmitting the adjusted model parameters to the second device. According to clause 56,
receiving, by the second device, the adjusted model parameters;
re-parameterizing the prediction model using the model parameters; and
The method further comprising generating additional future data values using the reparameterized prediction model. A method for prioritizing predictive model data streams, the method comprising:
Receiving, by a first device, a plurality of predictive model data streams, each of the predictive model data streams comprising a set of model parameters for a corresponding predictive model, each predictive model having a future data value of a data source. trained to predict -;
prioritizing, by the first device, priorities for each of the plurality of prediction model data streams;
selecting at least one of the predictive model data streams based on a corresponding priority;
parameterizing, by the first device, a predictive model using a set of model parameters included in the selected predictive model data stream; and
Predicting, by the first device, future data values of the data source using the parameterized prediction model. 59. The method of claim 58, wherein the selected at least one predictive model data stream is associated with high priority. 59. The method of claim 58, wherein selecting includes suppressing unselected predictive model data streams based on priorities associated with each unselected predictive model data stream. 59. The method of claim 58, wherein assigning priorities to each of the plurality of predictive model data streams includes determining whether each set of model parameters is anomalous. 59. The method of claim 58, wherein assigning priorities to each of the plurality of predictive model data streams includes determining whether each set of model parameters has changed from a previous value. 59. The method of claim 58, wherein the set of model parameters includes at least one vector. 64. The method of claim 63, wherein the at least one vector comprises a motion vector associated with a robot. 65. The method of claim 64, wherein the future data values include one or more future predicted positions of the robot. 59. The method of claim 58, wherein the predictive model predicts inventory levels of items, and the method further comprises:
based on the future data values, detecting an upcoming supply shortage of the item; and
The method further comprising taking steps to prevent shortage of the item. 59. The method of claim 58, wherein the predictive model is a behavior analysis model and the future data values represent predicted behavior of the entity. 59. The method of claim 58, wherein the predictive model is an augmented model and the future data values correspond to a non-operational sensor. 59. The method of claim 58, wherein the prediction model is a classification model and the future data values represent a predicted future state of a system comprising one or more sensor devices. 70. The method of claim 69, wherein the one or more sensor devices are RFID sensors associated with cargo and the future data values are indicative of future locations of the cargo. 70. The method of claim 69, wherein the one or more sensor devices are security cameras, and the data stream includes motion vectors extracted from video data captured by the security cameras. 70. The method of claim 69, wherein the one or more sensor devices are vibration sensors that measure vibrations produced by machines, and wherein the future data values indicate a potential need for maintenance of the machines. A method for updating one or more properties of one or more delivery digital twins, the method comprising:
Receiving a request to update one or more properties of one or more shipping digital twins;
Retrieving one or more shipping digital twins required to fulfill the request;
retrieving one or more dynamic models needed to fulfill the request;
selecting a data source from a set of available data sources based on one or more inputs of the one or more dynamic models;
Retrieving data from a selected data source;
using the retrieved data as one or more inputs to the one or more dynamic models to calculate one or more outputs; and
A method comprising updating one or more properties of the one or more shipping digital twins based on the output of the one or more dynamic models. 74. The method of claim 73, wherein the shipping digital twin is a digital twin of a smart container. 74. The method of claim 73, wherein the delivery digital twin is a digital twin of a delivery environment. 74. The method of claim 73, wherein the shipping digital twin is a digital twin of a shipping entity. 74. The dynamic model of claim 73, wherein the dynamic models include vibration, temperature, pressure, humidity, wind, rainfall, tides, storm surge, cloud cover, snowfall, visibility, radiation, audio, video, image, water level, quantum, flow, signal. Power, signal frequency, motion, displacement, speed, acceleration, light level, finance, cost, stock market, news, social media, revenue, operator, maintenance, productivity, asset performance, operator performance, operator reaction time, analyte concentration , a method of taking data selected from a set of biological compound concentration, metal concentration, and organic compound concentration data. 74. The method of claim 73, wherein at least one of the available data sources is an Internet of Things connected device, a machine vision system, an analog vibration sensor, a digital vibration sensor, a fixed digital vibration sensor, a three-axis vibration sensor, a single-axis vibration sensor, and an optical vibration sensor. A method selected from a set of vibration sensors. 74. The method of claim 73, wherein retrieving the one or more dynamic models comprises identifying the one or more dynamic models based on the one or more attributes indicated in the request and each type of the one or more shipping digital twins. , method. 74. The method of claim 73, wherein the one or more dynamic models are identified using a lookup table. A method of managing energy through a network of energy storage systems, the method comprising:
Communicating with one or more value chain network entities and associated data handling layers;
connecting one or more value chain network entities to a modular adaptive resource package to provide energy through the network of energy storage systems; and
Determining to use and store energy through the network of energy storage systems based on the use for energy identified in the data handling layer of the one or more value chain network entities. 82. The method of claim 81, wherein the data handling layers include at least one of adaptive intelligence systems, monitoring systems, data storage systems, interfaces, or connectivity facilities. 82. The method of claim 81, wherein the modular adaptive resource package providing energy through the network of energy storage systems includes three-dimensional printing of batteries, a battery energy storage system (BESS), various battery types, coordination processes, and distribution. energy grids, energy prices, energy storage technologies, energy-as-a-service, machine learning (ML) or artificial intelligence (AI) for energy optimization, ML or AI for automation, decentralized Energy for at least one of the following: ML or AI to match energy use/demand to energy production across the network, quantum computations, renewable energy, or technologies for slicing production, storage, or delivery. A method comprising the step of supplying. 82. The method of claim 81, wherein the modular adaptive resource package providing energy through a network of energy storage systems comprises energy storage technology. 85. The method of claim 84, wherein the energy storage technology includes lithium-ion batteries, flexible batteries, structural batteries, solid-state batteries, or flow batteries. 85. The method of claim 84, wherein the energy storage technology includes managing controlled vibration of dendrites to improve battery life. 82. The method of claim 81, wherein two or more of the value chain network entities are coupled to a modular adaptive resource package to provide energy through a network of energy storage systems to enable transactions between the value chain network entities. A method further comprising the step of connecting. 88. The method of claim 87, wherein connecting two or more of the value chain network entities to the modular adaptive resource package comprises tokenized energy assets purchased or sold in a decentralized marketplace.

Images