WO2023233786A1

WO2023233786A1 - Information processing system, information processing method, and program

Info

Publication number: WO2023233786A1
Application number: PCT/JP2023/012920
Authority: WO
Inventors: 久雄伊東
Original assignee: 株式会社シーズ
Priority date: 2022-06-02
Filing date: 2023-03-29
Publication date: 2023-12-07
Also published as: JP2023177649A

Abstract

The present invention addresses the problem of improving the convenience of management of facilities that are managed on the basis of data from a plurality of sensors or cameras.　In an information processing system including a server 1 that performs a prescribed process using input data, and one or more input cells 2 that provide at least a portion of the input data to the server 1, each input cell 2 acquires positional information of the input cell. The server 1 manages input cell position information and models, within a three-dimensional map of the facilities where the input cells 2 are located. Each input cell 2 performs relearning on the basis of information about the three-dimensional map managed by the server 1 and the positional information of the input cell 2. This solves the above-mentioned problem.

Description

Information processing system, information processing method, and program

The present invention relates to an information processing system, an information processing method, and a program.

Conventionally, there has been a technology that controls various devices based on parameters measured by various sensors in a premises (for example, a factory) to change the parameters and manufacture better products (for example, patent document (see 1).
Furthermore, in recent years, artificial intelligence (hereinafter referred to as "AI") technology has been developed. For example, AI capable of natural language processing, used in chatbots and the like, is also being developed.

Japanese Patent Application Publication No. 2013-111073

Conventionally, in order to detect and predict failures based on the operational status of various types of equipment used within the premises, software in a central management unit has been used to centrally detect, judge, and output alarms using various sensors. However, when an accident occurs and data (abnormal values) are sent from many sensors at once, the central monitoring device outputs an alarm to the user (for example, a supervisor) and activates emergency equipment in order to sequentially process the data. It takes a considerable amount of time to take measures such as fire extinguishing using sprinklers, which can lead to a delay in dealing with the situation and lead to a major disaster.
Therefore, there was a need for a system that could start responding more quickly and that could be used without building a large centralized system.

The present invention can improve the convenience of managing equipment that is managed based on data from a plurality of sensors or cameras.

To achieve the above object, an information processing system according to one embodiment of the present invention includes:
An information processing system including a central device that executes predetermined processing using input data, and one or more type 1 peripheral devices that provide at least a part of the input data to the central device,
The central device includes:
processing execution means that acquires one or more perceptual expression data as input data, executes a predetermined process using the input data, and outputs one or more perceptual expression data indicating the execution result of the process as output data;
Equipped with
Each of the one or more type 1 peripheral devices is
a model management means for storing and managing a model that inputs predetermined data and converts it into the perceptual expression data and outputs it in a predetermined storage medium;
Data output from a sensor that measures physical quantities in the real world or a camera that images a target area is acquired and input to the model, and the perceptual expression data output from the model is input to the central device. a conversion means outputting as at least part of the data;
Equipped with.

An information processing method and program according to one embodiment of the present invention correspond to the information processing system according to one embodiment of the present invention described above.

According to the present invention, it is possible to improve the convenience of managing equipment that is managed based on data from a plurality of sensors or cameras.

FIG. 1 is a schematic diagram illustrating an example of an outline of a service to which an "autonomous decentralized AI blockchain cell" is applied. 1 is a diagram showing a configuration example of an information processing system according to an embodiment of the present invention applied when providing the service shown in FIG. 1. FIG. 3 is a block diagram showing an example of the hardware configuration of a server in the information processing system shown in FIG. 2. FIG. 4 is a functional block diagram illustrating an example of a functional configuration of an information processing system including a server having the hardware configuration of FIG. 3. FIG. FIG. 2 is a diagram showing an example in which the service shown in FIG. 1 is used to manage devices located within a premises. FIG. 6 is a diagram showing an example of the flow of information processing for performing high-speed countermeasures in the premises of FIG. 5; FIG. 5 is a diagram illustrating an example of a flow for autonomously establishing processing contents in the input cell having the functional configuration of FIG. 4; FIG. 1 is a schematic diagram illustrating an example of an outline of a service to which an “autonomous decentralized AI blockchain audiovisual cell” is applied. 9 is a functional block diagram showing an example of a functional configuration of an information processing system that provides the service of FIG. 8. FIG. FIG. 8 is a diagram showing an example of using the service shown in FIG. 7 for managing devices located within a premises. FIG. 11 is a diagram illustrating an example of the flow of information processing for performing high-speed countermeasures in the premises of FIG. 10;

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

Note that in the following, when an image is simply referred to as an "image", it includes both a "moving image" and a "still image".
Furthermore, the "moving image" includes images displayed by each of the following first to third processing.
The first process is to continuously generate a series of still images over time for each movement of an object (for example, an image of a person or object to be imaged) in a two-dimensional image (2D image). This refers to the process of switching and displaying. Specifically, for example, two-dimensional animation, a so-called flipbook-like process, corresponds to the first process.
The second process is to set a motion corresponding to each movement of an object (for example, an image of a person or object to be imaged) in a stereoscopic image (image of a 3D model), and change the motion as time passes. This refers to the process of displaying images. Specifically, for example, three-dimensional animation corresponds to the second process.
The third process is a process in which images (that is, moving images) corresponding to the movements of objects (for example, images of natural people) are prepared, and the images are played over time.
Here, a "video (that is, a moving image)" is composed of images such as a plurality of frames and fields (hereinafter referred to as "unit images").

Hereinafter, the "autonomous decentralized AI blockchain cell" and the "autonomous decentralized AI blockchain audiovisual cell" which are the premise of the present invention will be explained using FIGS. 1 to 11.

The details will be described later, but here we will briefly explain the difference between the "autonomous decentralized AI blockchain cell" and the "autonomous decentralized AI blockchain audiovisual cell."
This service allows facilities (equipment, improvements, etc.) to respond to inputs from input cells (for example, input cells 2-1 and 2-2 in Figure 1) to output cells (for example, output cell 3-2 in Figure 1). 1) Block chain technology is used in which output is performed under predetermined control.

Specifically, an "autonomous decentralized AI blockchain cell" is one in which an input cell (for example, input cell 2-1 in Figure 1) is equipped with a sensor that measures a physical quantity in the real world (for example, temperature). Multiple units are installed within the facility and operate autonomously.
On the other hand, an "autonomous decentralized AI blockchain audiovisual cell" is one in which an input cell (for example, input cell 2-1 in Figure 8) is equipped with a camera and a microphone to image a predetermined area. There are multiple units installed within the building and they operate autonomously.

Hereinafter, the "autonomous distributed AI blockchain cell" will be explained using FIGS. 1 to 7, and the "autonomous distributed AI blockchain audiovisual cell" will be explained using FIGS. 8 to 11.
FIG. 1 is a schematic diagram illustrating an example of an overview of a service (hereinafter referred to as "this service") to which the "autonomous decentralized AI blockchain cell" is applied.
This service, to which "autonomous decentralized AI blockchain cell" is applied, operates the control device based on the input of data measured by multiple sensors by applying the information processing system shown in Figure 1. It provides a system for outputting the following.

The information processing system of this service shown in FIG. As shown in FIG. 1, it includes a "BC full node").
In addition, although the details will be described later, the input cells 2-1 and 2-2 and the output cell 3-1 have blockchain ultra-light nodes (hereinafter referred to as "BC ultra-light nodes" as shown in FIG. 1). It has the function of

The server 1 is an information processing device including a central AI 61. As will be described in detail later, the server 1 acquires as input data the measurement results of the sensors in the input cells 2-1 and 2-2 converted into perceptually expressed data, and based on the input data, outputs the output cells 3- The perceptual representation of the control instruction in step 1 is output as output data.

The input cells 2-1 and 2-2 are each equipped with a sensor and an autonomous distributed chip CP, and the measurement results of the sensors are converted into data perceptually expressed in the autonomous distributed chip CP and managed using the blockchain network BCN. .

The output cell 3-1 includes an autonomous distributed chip CP and an actuator (an example of a control device), acquires control instructions output by the server 1 in the autonomous distributed chip CP from the blockchain network BCN, and performs control based on the control instructions. Control the device (actuator).

The confirmation terminal 4 is an information processing device for confirming information managed using the blockchain network BCN. The person in charge of managing the equipment (factory, etc.) can use the confirmation terminal 4 to confirm the history of inputs and outputs that have not been tampered with.

Here, "perceptually expressed data" will be explained.
The raw data of the measurement results of the sensor is numerical data and is a physical quantity. On the other hand, humans can interpret the measurement results of such sensors and express the results using language or the like.
Specifically, for example, humans can perceive (recognize, interpret) raw data such as a temperature of 35 degrees and a humidity of 80% as hot and verbalize it. In this way, the perceived (recognized, interpreted) representation, rather than the raw data of sensor measurement results, is called a "perceptual representation."
Note that the form of "language" is an example of the form of perceptual expression. That is, for example, the data of the character string "hot" is data perceptually expressed in the form of the Japanese language. Further, for example, if identifiers are associated in advance in the information processing device, the data of the identifier corresponding to “hot” is also an example of perceptually expressed data that can be used in the information processing device. Furthermore, for example, video data of an action of looking up at the face with a hand can also be said to be data perceptually expressed in the form of a gesture.
Hereinafter, for ease of explanation, the explanation will be based on the assumption that "perceptual expression" is "language."

Additionally, as mentioned above, this service uses blockchain technology. Generally, the word blockchain can refer to distributed ledger technology or decentralized network. In other words, the word blockchain is a ambiguous word that includes a series of data called blocks connected like a chain, as well as technology and networks related to the data.
Therefore, hereinafter, the decentralized network that manages blockchain will be referred to as a "blockchain network" to distinguish it from a "blockchain," which is a series of data in which blocks are linked like a chain.
In other words, "blockchain" refers to various information (data itself, metadata, It is a series of data in which "blocks" containing hash values (data related to soundness verification, etc.) are connected like a chain.

First, the functions of various nodes in the blockchain network BCN in this service will be explained.

Blockchain network BCN is composed of multiple nodes, and at least one node exists on the cloud.
In the example of FIG. 1, each of the four BC full nodes 5-1 to 5-4 functions as a full node on the cloud. Here, a "full node" is an information processing device (node) that provides all the functions of a node in a blockchain, such as calculation processing functions related to block generation and storage functions of blockchain data itself.
The BC full nodes 5-1 to 5-4 form a blockchain network BCN that communicates with each other.

In the example of FIG. 1, the BC ultralight nodes provided in the two input cells 2-1 and 2-2 and the output cell 3-1 each function. Here, an "ultralight node" does not provide calculation processing functions related to block generation or storage functions for the blockchain data itself, but has an extremely simple function such as the function of sending and receiving data with the blockchain network BCN. This is an information processing device (node) that provides the functions of the section. Since ultralight nodes require fewer computational resources to implement, they are implemented as part or all of a chip or program that provides the above-mentioned functions. However, an ultralight node may also be implemented as a separate information processing device.

Although not shown, the nodes may include light nodes. Here, a "light node" is a node that does not function as a full node, but is responsible for part of the calculation processing function related to block generation and the storage function of the blockchain data itself.

The BC ultra light nodes provided in the input cells 2-1 and 2-2 and the output cell 3-1 are connected to the cloud via a dedicated line to communicate with the BC full nodes 5-1 to 5-4. conduct.
That is, in the example of FIG. 1, a total of 7 units, 4 BC full nodes 5-1 to 5-4 and 3 BC ultra-light nodes, function as each of the 7 nodes, so that the blockchain network BCN configured.

Here, a dedicated line is a communication line dedicated to a specific user.
For example, communication on a leased line is isolated from a network (eg, the Internet) that also includes untrusted information processing devices. That is, communications between information processing devices connected via a dedicated line are less likely to be wiretapped or intercepted by a malicious third party. That is, since this service is used via a dedicated line, it is provided with a low possibility of being wiretapped or intercepted by a third party. Note that the dedicated line does not need to be physically isolated from the Internet or the like. That is, for example, a virtual dedicated line using VPN (Virtual Private Network) technology may also be employed as the above-mentioned dedicated line.

An example of the configuration of the information processing system in this service to which the "autonomous decentralized AI blockchain cell" is applied has been described above using FIG. 1.
Hereinafter, the details of the flow of information processing in this service will be explained along steps ST11 to ST19 using FIG. 1.

In step ST11, numerical data indicating that "temperature is rising" and "humidity is rising" is acquired as a measurement result of the sensor of input cell 2-1. Specifically, for example, numerical data (digital signal) indicating that the temperature has increased from 20 degrees to 40 degrees over time, or numerical data (digital signal) indicating that the humidity has increased from 50% to 80% over time. ) is obtained.

In step ST12, the language conversion AI included in the autonomous distributed chip CP of the input cell 2-1 converts the digital signal into language. Specifically, for example, numerical data (digital signal) indicating that the temperature has increased from 20 degrees to 40 degrees over time, or numerical data (digital signal) indicating that the humidity has increased from 50% to 80% over time. ) is converted into the linguistic data ``hot''.

In step ST13, the BC ultra-light node possessed by the autonomous distributed chip CP of the input cell 2-1 manages the linguistic data "hot" using the blockchain network BCN. Specifically, for example, linguistic data such as "hot" is sent to the BC full node 5-1 as part of a transaction in blockchain technology to be managed. At this time, the BC ultralight node included in the autonomous distributed chip CP of the input cell 2-1 manages the encrypted language data using the blockchain network BCN.

Although not shown, the processes of steps ST11 to ST13 described above are also executed in the input cell 2-2.
Specifically, for example, data indicating that there are many heat sources is acquired as a measurement result of the sensor of the input cell 2-2. Specifically, for example, infrared thermography data is acquired as data indicating that there are many heat sources.
Further, for example, the language conversion AI included in the autonomous distributed chip CP of the input cell 2-2 converts data indicating that there are many heat sources into linguistic data that indicates "there are many people."
Further, for example, the BC ultralight node possessed by the autonomous distributed chip CP of the input cell 2-2 manages the encrypted linguistic data "There are many people" using the blockchain network BCN.

In step ST14, the server 1 decrypts and obtains the encrypted language data managed using the blockchain network BCN.
In step ST15, the server 1 uses the blockchain network BCN to manage the result of determining the content of the control instruction based on the acquired language data.
Specifically, for example, the server 1 uses the trained central AI 61 as an AI that performs natural language processing using input data such as ``hot'' and ``there are a lot of people.'' The language data of the control instruction "lower down" is generated as output data. The server 1 uses the blockchain network BCN to manage the generated control instruction to "lower the overall temperature more strongly." At this time, the server 1 manages the encrypted control instructions using the blockchain network BCN.

In step ST16, the BC ultralight node possessed by the autonomous decentralized chip CP of the output cell 3-1 obtains an encrypted control instruction managed using the blockchain network BCN.

In step ST17, the firmware included in the autonomous distributed chip CP of the output cell 3-1 decodes the encrypted control instruction. Then, the language conversion AI included in the autonomous distributed chip CP of the output cell 3-1 converts it into a digital signal as the specific control content of the control device.
Specifically, for example, the language conversion AI possessed by the autonomous decentralized chip CP of output cell 3-1 increases the cooling output of the air conditioner based on the control instruction to ``lower the overall temperature more strongly,'' and The signal is converted into a digital signal that allows the control device to operate in all wind directions. That is, for example, the language conversion AI included in the autonomous distributed chip CP of the output cell 3-1 outputs a control signal (digital signal) of digital data to operate an actuator that operates the air conditioner cooling.
Although not shown in the diagram, the BC Ultra Light node possessed by the autonomous decentralized chip CP of output cell 3-1 is also capable of controlling the operation of the control device that increases the cooling output of the air conditioner and makes it possible for the wind to flow in all directions using the blockchain network. It can be managed using BCN.

In step ST18, the actuator (an example of a control device) is driven in response to a digital data control signal (digital signal) for operating the actuator that operates the air conditioner cooling, such as "output improvement" and "output omnidirectional". As a result, the actuator (an example of a control device) operates in accordance with the control instruction from the central AI 61 to "lower the temperature overall."
As a result, based on sensor measurement results such as "temperature rise,""humidityrise," and "multiple heat sources," the air conditioner's cooling output is set to "improve output" and "output in all directions" (not shown), thereby improving the environment. Improve.

In step ST19, the confirmation terminal 4 can present the information managed using the blockchain network BCN to the manager of the facility (factory, etc.). That is, the manager of the facility (factory, etc.) can check the sensor measurement results, such as "temperature rise", "humidity rise", "many heat sources", etc., displayed on the confirmation terminal 4, "hot", "hot", "many heat sources", etc. Input data (linguistic data) such as "There are many people", output data (linguistic data) of control instructions such as "lower the temperature overall", "improve the output" of the air conditioner and make it "wind in all directions" The operation of the control device can be confirmed. Since these are managed using the blockchain network BCN, they are presented to the manager of the facility (factory, etc.) as a history of inputs and outputs that have not been tampered with.

With the configuration and operation described above, this service has the following features.
First, an existing AI trained in a language can be employed as the central AI 61.
That is, in this service, an existing AI that has been trained in a language can be used as the central AI 61, so overall development costs and operating costs can be reduced.
Specifically, when constructing such a system for new equipment, the sensors used and the prerequisite environments are different for each equipment, so the values that the sensor measurement results should take are also different. It is necessary to train the AI accordingly.
However, AI capable of natural language processing that generates input data (linguistic data) such as "hot" and "many people" and output data (linguistic data) of control instructions such as "lower the temperature overall" is Since it is used for many purposes, various developments are underway. Therefore, the current situation is that AI capable of such natural language processing is becoming more accurate and less expensive.
Therefore, this service can reduce overall development costs and operational costs.

Second, since AI processing is performed on the edge side, the processing load using the central AI 61 is reduced.
In other words, in this service, input cells 2-1 and 2-2 (edge side) perform processing to convert sensor measurement results into linguistic data (an example of perceptual expression), and output cell 3-1 (edge side) ), processing is performed to convert linguistic data (output data of control instructions) into digital signals as specific control contents of the control device.
In other words, the central AI 61 does not need to perform these processes, so its burden is reduced. In other words, it can be said that distributed processing of AI processing has been realized. Further, the effect of this distributed processing becomes more remarkable as the number of input cells and output cells on the edge side increases.

Third, this service uses the blockchain network BCN to guarantee data, so a safe AI system is realized.
For example, in a system that does not use the blockchain network BCN, by impersonating input cell 2-1 and inputting verbal data such as "temperature drop" to the central AI 61, an incorrect control instruction may be output. .
In this service, since data is guaranteed using the blockchain network BCN, such impersonation of the input cell 2-1 becomes impossible, and a safe AI system is realized.

The outline of this service has been explained above using FIG. 1.
The information processing system applied when providing the service shown in FIG. 1 of this service will be described below with reference to FIGS. 2 to 4.
FIG. 2 is a diagram showing a configuration example of an information processing system according to an embodiment of the present invention that is applied when providing the service shown in FIG. 1.

That is, the configuration example of the information processing system shown in FIG. 2 is a more general system configuration of the information processing device of this service shown in FIG.

The server 1 is an information processing device managed by a service provider S. The server 1 includes input cells 2-1 to 2-N (N is an integer value of 1 or more), output cells 3-1 to 3-M (M is an integer value of 1 or more independent of N), and a confirmation terminal 4. , BC full nodes 5-1 to 5-L (L is an integer value of 1 or more independent of N and M), and executes various processes to realize this service.

The input cells 2-1 to 2-N are information processing devices including one or more sensors and an autonomous distributed chip CP. Furthermore, when one of the N input cells 2-1 to 2-N is illustrated and explained, "input cell 2-p" (p is an integer value of 1 or more and N or less) is used.

The output cells 3-1 to 3-M are information processing devices including one or more control devices and an autonomous distributed chip CP. Furthermore, when one of the M output cells 3-1 to 3-M is illustrated and explained, "output cell 3-q" (q is an integer value of 1 or more and M or less) is used.

The confirmation terminal 4 is an information processing device operated by a person in charge of management of a facility (factory, etc.) to confirm information managed using the blockchain network BCN. Note that although one confirmation terminal 4 is illustrated in FIG. 2, the number of confirmation terminals 4 is arbitrary.

The BC full nodes 5-1 to 5-L are information processing devices (nodes) that provide all functions as nodes in the blockchain, such as calculation processing functions related to block generation and storage functions of the blockchain data itself.
The BC full nodes 5-1 to 5-L in the example of FIG. 2 exist on cloud C.

FIG. 3 is a block diagram showing an example of the hardware configuration of a server in the information processing system shown in FIG. 2.

The server 1 includes a CPU (Central Processing Unit) 11, a ROM (Read Only Memory) 12, a RAM (Random Access Memory) 13, a bus 14, an input/output interface 15, an input unit 16, and an output unit. Part 17 and , a storage section 18, a communication section 19, and a drive 20.

The CPU 11 executes various processes according to programs recorded in the ROM 12 or programs loaded into the RAM 13 from the storage section 18 .
The RAM 13 also appropriately stores data necessary for the CPU 11 to execute various processes.

The CPU 11, ROM 12, and RAM 13 are interconnected via a bus 14. An input/output interface 15 is also connected to this bus 14 . An input section 16 , an output section 17 , a storage section 18 , a communication section 19 , and a drive 20 are connected to the input/output interface 15 .

The input unit 16 includes, for example, a keyboard, and inputs various information.
The output unit 17 includes a display such as a liquid crystal display, a speaker, and the like, and outputs various information as images and sounds.
The storage unit 18 is composed of a DRAM (Dynamic Random Access Memory) or the like, and stores various data.
The communication unit 19 communicates with other devices (for example, input cells 2-1 to 2-N, output cells 3-1 to 3-M, confirmation terminal 4, BC full nodes 5-1 to 3-M in FIG. 2) via a network including the Internet. 5-L).

A removable medium 31 made of a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is appropriately installed in the drive 20. The program read from the removable medium 31 by the drive 20 is installed in the storage unit 18 as necessary.
Further, the removable medium 31 can also store various data stored in the storage section 18 in the same manner as the storage section 18 .

Although not shown, the input cells 2-1 to 2-N, output cells 3-1 to 3-M, confirmation terminal 4, and BC full nodes 5-1 to 5-L in FIG. 2 also have the hardware shown in FIG. It can have basically the same configuration as the hardware configuration. Therefore, a description of the hardware configurations of the input cells 2-1 to 2-N, the output cells 3-1 to 3-M, the confirmation terminal 4, and the BC full nodes 5-1 to 5-L will be omitted.
However, as shown in FIG. 1, the input cells 2-1 to 2-N are equipped with a sensor as an input section. Further, the input cells 2-1 to 2-N each have a part or all of a CPU, ROM, RAM, etc. as an autonomous distributed chip CP.
Similarly, the output cells 3-1 to 3-M are equipped with a control device (for example, an actuator) as an output section, as shown in FIG. Further, the output cells 3-1 to 3-M have a part or all of a CPU, ROM, RAM, etc. as an autonomous distributed chip CP.

Through cooperation between various hardware and various software of the server 1 shown in FIG. 3, various processes can be executed. As a result, the above-mentioned service can be provided.
The functional configuration of the information processing system of FIG. 2, including the server 1 of FIG. 3, will be described below.

FIG. 4 is a functional block diagram showing an example of the functional configuration of an information processing system including a server with the hardware configuration of FIG. 3.

As shown in FIG. 4, a processing execution unit 51 functions in the CPU 11 of the server 1. Furthermore, the storage unit 18 stores a model of the central AI 61.

In the input cell 2-p, a model management unit 71, a language conversion AI 72, a relearning execution unit 73, and a sensor 74 function as an autonomous distributed chip CP. The input cell 2-p has a sensor 74 as an input section. Furthermore, a language conversion AI model 75 is stored in the storage section of the input cell 2-p.

In the output cell 3-q, a model management unit 81, an inverse language conversion AI 82, a control unit 83, and a relearning execution unit 84 function as an autonomous distributed chip CP. The output cell 3-1 has an actuator 85 as an output section. Furthermore, an inverse language conversion AI model 86 is stored in the storage section of the output cell 3-q.

The processing execution unit 51 acquires input language data as input data from one or more input cells 2-p, executes a predetermined process using the input data, and outputs one or more output languages indicating the execution results of the process. Output the data as output data.
The processing execution unit 51 obtains input data stored using the blockchain network BCN.

The model management unit 71 receives a signal indicating the measurement result of the sensor 74, converts it into language data, and outputs the language conversion AI model 75. The model management unit 71 stores and manages a language conversion AI model 75 in a storage unit.
Specifically, for example, the model management unit 71 converts and outputs linguistic data such as "temperature rise", "humidity rise", "many people", etc. based on the results of snooping measurements such as temperature, humidity, and thermography. do.

The language conversion AI 72 acquires a signal indicating a predetermined physical quantity output from a sensor 74 that detects a predetermined physical quantity in the real world, inputs it to a language conversion AI model 75, and converts the language data output from the model into a language conversion AI model 75. , output as at least a part of the input language data of the server 1.

The relearning execution unit 73 executes relearning to update the model.

The BC ultralight node of the input cell 2-p stores at least a portion of the post-input language data output from the language conversion AI 72 using the blockchain network BCN.

A model management unit 81 stores and manages an inverse language conversion AI model 86 that inputs language data, converts it into a predetermined physical quantity, and outputs it in a predetermined storage medium.

The BC ultralight node of the output cell 3-q uses the blockchain network BCN to obtain the stored output language data and provides it to the inverse language conversion AI 82.
The inverse language conversion AI 82 acquires at least a part of the one or more linguistic expression data that constitutes the output data output from the server 1, inputs it to the inverse language conversion AI model 86, and causes the model to output a signal indicating a physical quantity. .

The control unit 83 controls the actuator 85 by inputting a signal indicating the physical quantity output from the model to the actuator 85 as an instruction signal.

The relearning execution unit 84 executes relearning to update the model.

Note that an example of relearning by the relearning execution unit 73 and the relearning execution unit 84 will be described later.
With such a functional configuration, the information processing system of this service can execute each process described using FIG. 1 and the like.

Furthermore, the information processing system of this embodiment can be utilized as shown below.
Hereinafter, an example in which the present information processing system is applied to monitor a plurality of devices arranged in a premises will be described using FIGS. 5 and 6.
FIG. 5 is a diagram showing an example in which the service shown in FIG. 1 is used to manage devices located within a premises.
In the example of FIG. 5, four devices A1 to A4 are arranged within the premises. That is, for example, the devices A1 to A4 are devices on a manufacturing line located within the premises of a factory.

An input cell 2-1 is arranged in the device A1. This shows, for example, that a certain event related to the device A1 on the production line is being measured by the sensor of the input cell 2-1. Specifically, for example, the sensor measures an arbitrary physical quantity, such as the temperature at a predetermined location in the device A1 or the weight of the material put into the device A1.

Furthermore, input cells 2-2 and 2-3 are arranged in the device A2. This shows how two physical quantities of one device A2 are measured by the sensors of two input cells 2-2 and 2-3. That is, in the device A1, the physical quantity of one device A1 was measured by the sensor of one input cell 2-1, but a plurality of input cells can be arranged for one device.

Furthermore, an input cell 2-4 is arranged inside the device A3. This shows that the input cell 2-4 is built into the device A3, and the sensor of the input cell 2-4 measures the physical quantity of the device A3.

Furthermore, input cells 2-5 and 2-6 are arranged inside the device A4. This shows that the two input cells 2-5 and 2-6 are built into one device A4, and the sensors of the two input cells 2-5 and 2-6 measure the physical quantities of the device A4. It shows.

Then, each input cell 2-1 to 2-6 converts the data measured by the sensor into linguistic data and sends it to the edge server EDS. In this way, data of each of the plurality of devices A1 to A4 within the premises is collected.

This makes it possible to detect and predict failures based on the operating status of various devices used within the premises (for example, manufacturing robots (lines), room temperature control, lighting control, etc.). Furthermore, unmanned monitoring of the premises (for example, factories, hospitals, apartments, etc.) shown in FIG. 5 becomes possible.
As mentioned above, in conventional monitoring systems, when an accident occurs and data (abnormal values) are sent from many sensors at once, the central monitoring device processes the data sequentially, so the user (for example, a monitoring staff) It sometimes took a considerable amount of time to take measures such as issuing alarms and operating emergency equipment (for example, extinguishing fires with sprinklers).

On the other hand, in this information processing system, each input cell is converted into language data, so processing is distributed and the system can handle the system at high speed as a whole.

Furthermore, by processing without using control instructions from the central AI 61, it is possible to take action even faster.
FIG. 6 is a diagram illustrating an example of the flow of information processing when performing faster response in the premises of FIG. 5.
In the example shown in FIG. 6, language data is provided from input cell 2-p to output cell 3-q. That is, the language data from the input cell 2-p may be provided to the output cell 3-q without going through the central AI 61.
It is preferable that the input cell 2-p and the output cell 3-q have the following basic logic, especially in an "autonomous decentralized AI blockchain cell".

First, it is preferable that the input cell 2-p and the output cell 3-q have "individuality establishing logic."
That is, the "individuality establishment logic" is a logic processing algorithm that is individually prepared and determined for each input cell 2-p and output cell 3-q, depending on the sensor measurement object, etc.
That is, for example, drive devices such as motors are subject to manufacturing variations and individual deterioration over time. Specifically, for example, even if the motors are of the same type, some motors vibrate and abnormally overheat at 1000 rpm (rotations per minute), while others rotate normally at 2000 rpm.
Each input cell 2-p, which quickly generates various elements (physical quantities) using each sensor attached to each motor, individually learns the characteristics of the motor it manages, and determines what is normal for itself. You can set the limiter yourself depending on the line. The input cell 2-p can provide verbal data such as "abnormal motor rotation speed" when the limiter set individually by the input cell 2-p is exceeded.
Note that the individuality establishment logic can be similarly applied to the stable operating temperature, etc. of control boards other than motors.

Second, it is preferable that the input cell 2-p and the output cell 3-q have "other comparison learning logic".
In other words, "other-comparison learning logic" is an algorithm that grasps the uniqueness of the device it manages compared to other devices based on the results of exchanging information with other cells. be. Specifically, the input cell 2-p and the output cell 3-q determine the characteristics of the devices they manage (targets measured by sensors and targets controlled via control devices) relative to other devices. By sharing data with other cells, you can understand whether there is a cell. This allows output information to be corrected using shared data.
That is, the language conversion AI model 75 of the input cell 2-p executes a relearning process by the relearning execution unit 73 based on language data from other input cells. That is, for example, to explain using the example of FIG. 1, when the linguistic data "There are few people" is output, a large number of input cells including other input cells 2-2 output the linguistic data "There are many people". Suppose you are outputting . At this time, it is assumed that the input cell 2-p outputs linguistic data indicating that there are "few people" even though the areas monitored by the sensors are the same. In such a case, the relearning execution unit 73 of the input cell 2-p can perform a relearning process to more easily output "there are many people".
In this way, the input cell 2-p can learn based on (by comparison with) the language data output by other input cells (others).

Thirdly, it is preferable that the input cell 2-p and the output cell 3-q have "awareness/intent communication logic."
The "consciousness/intent communication logic" is a logic algorithm in which the language conversion AI of the input cell 2-p simultaneously outputs human words such as consciousness and intention in addition to the digital data of the normal sensor output.
Specifically, for example, the language conversion AI 72 of the input cell 2-p outputs language data such as "It's a little hot,""I feel a lot of vibration," and "This is a dangerous situation." By having the central AI 61, which is an existing language input AI, directly grasp this language data, it is possible to ask the central AI 61 to think and make decisions.
That is, when an emergency is required, the edge server EDS in FIG. 6 notifies the off-premises server 1 (server 1 in FIG. 4, etc., not shown in FIG. 6), and at the same time waits for the higher-level judgment result by the server 1 in the suburbs. You can take corrective action without any problems.

Fourthly, it is preferable that the input cell 2-p and the output cell 3-q have "self-handling logic."
If there is no need to ask the central monitoring device for judgment, for example, if a clear abnormality is detected, this information processing system can take action without waiting for a response from the edge server EDS or the server 1 (not shown). I can do it. Specifically, for example, if the input cell 2-p outputs linguistic data saying "fire outbreak" instead of "temperature rise", that linguistic data is provided to the output cell 3-q and causes an alarm to sound. or activate local sprinklers.
In other words, in this information processing system, if the central AI 61 is likened to the human brain, it is possible to respond without waiting for the brain to make a decision, which can be called a reflex.
This realizes information processing for high-speed countermeasures.

Fifth, it is preferable that the input cell 2-p and the output cell 3-q have "ethical logic."
That is, according to the logic of ethics, the edge server EDS and the server 1 (not shown) outside the premises can make total predictions and judgments, output language data for control instructions, and provide education to each cell. .
That is, the language conversion AI model 75 of the input cell 2-p executes a relearning process by the relearning execution unit 73 based on the output language data etc. from the central AI 61. Through such learning, the input cell 2-p can output higher level linguistic data such as "fire outbreak" instead of "temperature rise". This allows the input cell 2-p to make individual decisions and take instantaneous countermeasures depending on the processing stage of the input cell 2-p, thereby making it possible to prevent major accidents.
That is, this realizes information processing when taking high-speed countermeasures.

In this way, the input cell 2-p and output cell 3-q of this information processing system realize the concept of identity through IT, and can be called "autonomous decentralized AI blockchain cells." It is.

Furthermore, the input cell 2-p and output cell 3-q of this information processing system, as "autonomous decentralized AI blockchain cells", autonomously judge the processing content and request processing content instructions from the central AI 61. can do.
FIG. 7 is a diagram showing an example of a flow for autonomously establishing processing contents in the input cell having the functional configuration of FIG. 4.

As shown in step ST21, the input cell 2-1 shown in FIG. 7 has acquired temperature information from the first sensor and humidity information from the second sensor.
At this time, the autonomous distributed chip of the input cell 2-1 autonomously recognizes the "mission" of the input cell 2-1 based on information obtained from the sensor.
Here, the mission refers to the processing content that the input cell 2-1 should perform, and the specific processing content such as what kind of language data should be output from temperature information and humidity information. It refers to
That is, for example, input cell 2-1 identifies that the acquired data is temperature and humidity information based on the data format, and determines that its mission is to determine the temperature (humidity) environment. .
Similarly, in the input cell 2-2, as a result of acquiring heat sensation (thermography data, etc.) from the sensor, the input cell 2-2 has the mission of detecting the density of the heat source (especially the density of people in the example of Fig. 1). ).
In this way, the input cells 2-1 and 2-2 can autonomously recognize their own missions.

As shown in step ST22, the input cells 2-1 and 2-2 transmit their own missions to the central AI 61.

As shown in step ST23, the central AI 61 can make an integrated decision on the missions transmitted from each cell and determine whether the mission of each cell is appropriate.
The central AI 61 determines whether the mission autonomously recognized by each cell is appropriate. Then, if the mission autonomously recognized by each cell is not valid, the central AI 61 transmits (correct) mission information.
As a supplement here, the central AI 61 is capable of performing various judgment processes in addition to the judgment process for outputting output language data from input language data in the above description. Specifically, for example, an overall image of the equipment, information on the arrangement schedule of the input cell 2-p, etc. are stored in advance, and processing such as integrated judgment with information from the input cell can be performed.

As shown in step ST24, the autonomously recognized mission is compared with the (correct) mission information transmitted from the central AI 61 to finalize the cell's own mission.
The input cells 2-1 and 2-2 then execute subsequent processing based on the finally determined mission.

To summarize the above, the input cells 2-1 and 2-2 can recognize their own missions based on information from the sensors.
As a result, the individuality of the input cells 2-1 and 2-2 is established, and they operate as cells dedicated to the sensors.
In addition, when input cells 2-1 and 2-2 recognize the mission when information from sensors and conjunctions and sensors is acquired, when producing input cells 2-1 and 2-2, settings must be made individually. It is possible to ship the product in a neutral state without having to do so. This allows cost reduction during production of the input cells 2-1 and 2-2.

The embodiments of the "autonomous decentralized AI blockchain cell" have been described above, but the "autonomous decentralized AI blockchain cell" is not limited to the above embodiments, and is within the range that can achieve the purpose of the present invention. Modifications, improvements, etc. may be made as appropriate.

In the above-described embodiment, the input cell 2-p generates language data from the measured value of the sensor, but the present invention is not limited thereto.
In other words, for example, if a sensor has the function of replacing a "language" for humans with a "language" for machines, and a sensor detects "temperature rise" in human language, an identifier has been previously associated with the information processing device. , it may be converted into an identifier corresponding to "hot".
This makes it possible for this information processing system to employ the results of checks made by the human eye as input language data.

Also, for example, the sensor may be a microphone or a camera. That is, for example, the language conversion AI 72 may specify a person from the voice acquired by the microphone and output it as language data, or may specify the emotion of the person and output it as language data. For example, the language conversion AI 72 may convert human gestures in a video captured by a camera into language data.

As mentioned above, the "autonomous decentralized AI blockchain cell" uses sensor functions to process sensors such as temperature, humidity, pressure, vibration, altitude, sound pressure, brightness, ultrasonic waves, voltage, and current, as well as blockchain-related processes. A unit of executable information processing equipment. Note that although the information processing device is used, the present invention is not limited to a personal computer or a server device, and can be implemented as a chip that is an electronic component. This makes it possible to easily incorporate it into various conventional devices.

Next, an embodiment including an "autonomous decentralized AI blockchain audiovisual cell", that is, an input cell equipped with a camera, will be described below with reference to FIGS. 8 to 11.
In addition, regarding the input cell equipped with a camera, in order to distinguish it from the above-mentioned "input cell 2-p", it is particularly referred to as "input cell 2-r" (r is an integer value of 1 or more and N or less and other than p). It is called.

In other words, as conventional unmanned monitoring of premises (factories, hospitals, apartments, etc.), that is, monitoring without people staying in the premises, security monitoring of the premises and monitoring of various devices used within the premises are possible. It is being done. Specifically, captured images are recorded on a recorder and monitored by humans to detect suspicious persons, detect failures, and predict failures.
Therefore, even in cases where instantaneous judgment is required, such as a sudden incident or accident, the human intervention may delay the response and lead to a major disaster.

For example, conventionally, in order to detect suspicious persons wandering around or people who have committed a dangerous act doing the same thing again, the person in charge (human) has to look at images from surveillance cameras and make judgments. Methods such as checking images recorded on a recorder are used.
In addition, when detecting events such as suspicious people wandering around even though entry is prohibited to anyone but those involved, or a person who has committed a dangerous act doing the same thing again, The person in charge would either look at images from a surveillance camera and make a decision, or they would look at recordings after a report such as the presence of a suspicious person.
In this way, when the person in charge makes a decision, there is a possibility that the response will be delayed due to checking the images of each surveillance camera, etc., and the degree of damage caused by the accident may increase.

Additionally, automatic fire detection devices that utilize heat sensors, smoke sensors, and the like have conventionally existed. However, such conventional automatic fire detection devices detect a fire after it has occurred. In other words, conventional automatic fire detection devices cannot detect abnormalities before a fire occurs, and cannot take preventive measures based on predictions of the occurrence of a fire.
For example, in the event of an accident such as a fire, there are automatic detection devices that use heat sensors, smoke sensors, etc., but although it is possible to detect a fire after it has started, it is not possible to detect an abnormality in advance. It was impossible to do.
Therefore, there was a need for a system that could start responding more quickly and that could be used without building a large centralized system.

The "autonomous decentralized AI blockchain audiovisual cell" contributes to providing a system that more appropriately processes data of images captured by multiple cameras.

FIG. 8 is a schematic diagram illustrating an example of the outline of this service to which the "autonomous decentralized AI blockchain audiovisual cell" is applied.
This service to which the "autonomous decentralized AI blockchain audiovisual cell" is applied uses the information processing system shown in Figure 8 to control the control device based on the input of image data captured by multiple cameras. It provides a system for outputting operations.
In the explanation of FIG. 8, the input cells equipped with cameras will be explained using the symbols 2-3 and 2-4 to distinguish them from the input cells in the explanation of FIG.

The information processing system of this service shown in FIG. 8 includes a server 1, input cells 2-3 and 2-4, output cell 3-1, confirmation terminal 4, and blockchain full nodes 5-1 to 5-4 (hereinafter As shown in FIG. 1, it includes a "BC full node").
In addition, although the details will be described later, the input cells 2-3 and 2-4 and the output cell 3-1 have blockchain ultra-light nodes (hereinafter referred to as "BC ultra-light nodes" as shown in FIG. 1). It has the function of

The server 1 is an information processing device including a central AI 61. As will be described in detail later, the server 1 converts the image data captured by the cameras CAM-1 and CAM-2 in the input cells 2-3 and 2-4 into perceptually expressed data as input data. A perceptual representation of the control instruction in the output cell 3-1 based on the input data is output as output data.

The input cells 2-3 and 2-4 are equipped with cameras CAM-1 and CAM-2 and an autonomous distributed chip CP, respectively, and convert the image data captured by the cameras into the data perceptually expressed in the autonomous distributed chip CP. It will be converted into and managed using the blockchain network BCN.

Here, a specific example of "perceptually expressed data" in an image captured by a camera will be described.
Data (raw data) of images captured by a plurality of cameras is, for example, numerical data of colors in each of pixels arranged two-dimensionally in a still image. On the other hand, humans can express the results of interpreting such images using language or the like.
Specifically, for example, in an image where the first point is captured, even though the image of a certain man has never been included up to the first time (for example, over several months), If the image is included for the first time at the first time, a person viewing (viewing) the image can perceive (recognize, interpret) that there is a man who is not usually present at the first location at the first time, and verbalize it. In this way, the perceived (recognized, interpreted) representation, rather than the image data (raw data), is called a "perceptual representation."
Note that the form of "language" is an example of the form of perceptual expression. That is, for example, the data of the character string "fire" is data perceptually expressed in the form of the Japanese language. For example, if identifiers are associated in advance in the information processing device, the data of the identifiers corresponding to “first point” and “fire” are also examples of perceptually expressed data that can be used in the information processing device. . Furthermore, for example, data of a still image of a flame icon can also be said to be data perceptually expressed in the form of an icon corresponding to "fire".

First, the functions of various nodes in the blockchain network BCN in this service shown in FIG. 8 are basically the same as in FIG. 1.
Furthermore, in the example of FIG. 8, the functions of the BC ultralight nodes provided in the two input cells 2-3 and 2-4 and the output cell 3-1 are basically the same. Therefore, the explanation will be omitted.

An example of the configuration of the information processing system in this service to which the "autonomous decentralized AI blockchain cell audiovisual cell" is applied has been described above using FIG. 8.
Hereinafter, using FIG. 8, details of the flow of information processing in this service will be explained along steps ST21 to ST29.

In step ST21, it is assumed that the camera CAM-1 provided in the input cell 2-3 is capturing an image of a first point at a first time. Then, as the data of the image captured by the camera CAM-1, data of "an image including an image of a certain man" is acquired.

In step ST22, the language conversion AI included in the autonomous distributed chip CP of the input cell 2-3 converts image data (digital signal) into language. For example, assume that the image of a man included in the image has never been captured before the first time. In such a case, the image data (digital signal) is converted into verbal data that says "There is a man who is not usually present at the first location at the first time."

In step ST23, the BC ultra-light node possessed by the autonomous decentralized chip CP of the input cell 2-3 uses the blockchain network BCN to transmit the linguistic data "There is a man who is usually absent at the first location at the first time". Let them manage it. Specifically, for example, by sending language data such as "There is a man who is usually absent at the first location at the first time" to the BC full node 5-1 as part of a transaction in blockchain technology, Let them manage it. At this time, the BC ultralight node included in the autonomous distributed chip CP of the input cell 2-3 manages the encrypted language data using the blockchain network BCN.
Note that at this time, the characteristics of the man (height, physique, characteristics of clothing, etc.) may be managed as appropriate.
Further, the data of the captured image itself may also be appropriately managed using the blockchain network BCN. Specifically, for example, metadata of image data is managed using the blockchain network BCN. Then, the image data itself is appropriately encrypted and divided, and managed using IPFS or the like.

Further, although not shown, the same processing as steps ST21 to ST23 described above is executed in the input cell 2-4 as well.
Specifically, for example, it is assumed that the camera CAM-2 provided in the input cell 2-4 images the second point at a second time after the first time. Then, data of "an image including an image of a certain man" is acquired as data of the image captured by the camera CAM-2.
Then, the language conversion AI included in the autonomous distributed chip CP of the input cell 2-4 converts the image data (digital signal) into language. For example, assume that the image of a man included in the image has never been captured before the first time. In such a case, the image data (digital signal) is converted into verbal data that says, "There is a man who is not usually present at the second location at the second time."
Then, the BC ultralight node possessed by the autonomous decentralized chip CP of input cell 2-4 sends the encrypted linguistic data "There is a man who is usually absent at the second location at the second time" to the blockchain network BCN. be used and managed.

In step ST24, the server 1 decrypts and obtains the encrypted language data managed using the blockchain network BCN.
In step ST25, the server 1 uses the blockchain network BCN to manage the result of determining the content of the control instruction based on the acquired language data.
Specifically, for example, the server 1 inputs linguistic data such as "At the first time, there is a man who is not usually seen at the first location" and "At the second time, there is a man who is not usually seen at the second location" as input data. Using the trained central AI 61 as an AI that performs natural language processing, language data of a control instruction to "warn the man at the second location" is generated as output data. The server 1 manages the generated control instruction "warn the man at the second location" using the blockchain network BCN. At this time, the server 1 manages the encrypted control instructions using the blockchain network BCN.

In step ST26, the BC ultralight node possessed by the autonomous decentralized chip CP of the output cell 3-1 acquires the encrypted control instruction managed using the blockchain network BCN.

In step ST27, the firmware included in the autonomous distributed chip CP of the output cell 3-1 decodes the encrypted control instruction. Then, the language conversion AI included in the autonomous distributed chip CP of the output cell 3-1 converts it into a digital signal as the specific control content of the control device.
Specifically, for example, the language conversion AI possessed by the autonomous decentralized chip CP of the output cell 3-1 is configured to ``send an alarm device around the second point'' based on the control instruction to ``warn the man at the second point.'' It is converted into a digital signal that causes a control device (actuator, control unit of an alarm device, etc.) to operate so as to cause a warning. That is, for example, the language conversion AI included in the autonomous distributed chip CP of the output cell 3-1 outputs a digital data control signal (digital signal) for causing an alarm device to issue.
Although not shown in the figure, the BC ultra light node possessed by the autonomous decentralized chip CP of the output cell 3-1 also controls the operation of the control device to "set off the alarm device around the second point" using the blockchain network BCN. can be used and managed.

In step ST28, the actuator (an example of a control device) is driven in response to a digital data control signal (digital signal) for "setting off a warning device" around the second point. As a result, the actuator (an example of a control device) operates in accordance with the control instruction from the central AI 61 to "warn the man at the second date."
As a result, based on the data of the image captured by the camera, such as "There is a man who is not usually seen at the first location at the first time" and "There is a man who is not usually seen at the second location at the second time", etc. An alarm device around a second point (not shown) is activated to issue a warning to the man.

In step ST29, the confirmation terminal 4 presents the information managed using the blockchain network BCN to the manager of the facility (factory, etc.). That is, the manager of the facility (factory, etc.) displays the data of the images at the first time and the second time, which were captured by the cameras at the first and second points, presented on the confirmation terminal 4. Input data (linguistic data) such as "There is a man who is usually not at the first location at the first time", "There is a man who is usually not at the second location at the second time", "Warning the man at the second location", etc. It is possible to confirm the output data (language data) of the control instruction "" and the operation of the control device that causes the "warning device to sound" around the second point. Since these are managed using the blockchain network BCN, they are presented to the manager of the facility (factory, etc.) as a history of inputs and outputs that have not been tampered with.

With the configuration and operation as described above, this service has the following features in the same way as explained in the explanation of FIG. 1 above.
First, an existing AI trained in a language can be employed as the central AI 61.
Therefore, this service can reduce overall development costs and operational costs.

Second, since AI processing is performed on the edge side, the processing load using the central AI 61 is reduced.
That is, the central AI 61 does not need to process perceptual expression (verbalization), so its burden is reduced. In other words, it can be said that distributed processing of AI processing has been realized. Further, the effect of this distributed processing becomes more remarkable as the number of input cells and output cells on the edge side increases.

Third, this service uses the blockchain network BCN to guarantee data, so a safe AI system is realized.
As a result, this service guarantees data using the blockchain network BCN, making it impossible to impersonate the input cell 2-3, thereby realizing a safe AI system.

In the above example, the linguistic AI outputs that there was a man who is not usually present at a certain point at a certain time. However, the linguistic AI can output various kinds of outputs. Specifically, for example, "sparks are flying at the first location (from the installed device)", "smoke is occurring at the first location", "a fire is occurring at the first location", A perceptual expression to that effect may be output.

The outline of this service to which the "autonomous decentralized AI blockchain cell" is applied has been explained above using FIG. 8.
The configuration of the information processing system applied when providing the service shown in Figure 8 of this service to which the "Autonomous Decentralized AI Blockchain Cell" is applied is basically the same as that shown in Figures 2 and 3. The same is true.

However, as shown in FIG. 1, the input cell 2-r is equipped with a camera CAM-p as an input section. Further, the input cell 2-r has a part or all of a CPU, ROM, RAM, etc. as an autonomous distributed chip CP.

Through cooperation between various types of hardware and various types of software of the server 1, various types of processing can be executed. As a result, the above-mentioned service can be provided.
The functional configuration of the information processing system that provides the services shown in FIG. 8 will be described below.

FIG. 9 is a functional block diagram showing an example of the functional configuration of an information processing system that provides the services shown in FIG. 8.

As shown in FIG. 9, a processing execution unit 51 functions in the CPU 11 of the server 1. Furthermore, the storage unit 18 stores a model of the central AI 61.

In the input cell 2-r, a model management unit 71, a language conversion AI 72, and a relearning execution unit 73 function as an autonomous distributed chip CP. The input cell 2-r has a camera CAM-p as an input section. Furthermore, a language conversion AI model 75 is stored in the storage section of the input cell 2-r.

In the output cell 3-q, a model management unit 81, an inverse language conversion AI 82, a control unit 83, and a relearning execution unit 84 function as an autonomous distributed chip CP. The output cell 3-q has an actuator 85 as an output section. Furthermore, a language conversion AI model 86 is stored in the storage section of the output cell 3-q.

The processing execution unit 51 acquires input language data as input data from one or more input cells 2-r, executes a predetermined process using the input data, and outputs one or more output languages indicating the execution results of the process. Output the data as output data.
The processing execution unit 51 obtains input data stored using the blockchain network BCN.

The model management unit 71 stores and manages a language conversion AI model 75 that inputs image data captured by the camera CAM-p, converts it into language data, and outputs it in a storage unit.
Specifically, for example, if an image of a man that has never been imaged is included, the model management unit 71 converts it into linguistic data such as "There is a man who normally does not exist at the first point at the first time". and output it.

The language conversion AI 72 acquires image data of the target area output from the camera CAM-p that images the target area, inputs it to the language conversion AI model 75, and sends the language data output from the model to the server. output as at least a part of the input language data of 1.
The processing after being converted into language by the language conversion AI 72 is basically the same. Therefore, the explanation will be omitted.
Hereinafter, an example in which the present information processing system is applied to monitor a plurality of devices arranged in a premises will be described using FIGS. 10 and 11.
FIG. 10 is a diagram showing an example in which the service shown in FIG. 1 is used to manage devices located within a premises.
In the example of FIG. 10, two devices A1 and A2 are located within the premises. That is, for example, the devices A1 and A2 are devices on a manufacturing line located within the premises of a factory.

An input cell 2-1 is arranged in the device A1. This shows, for example, that the camera CAM-1 of the input cell 2-1 is capturing an image of the device A1 on the production line and its surroundings as a target area. Specifically, for example, the sensor captures an image of the device A1 as an image in which the appearance of the device A1, a predetermined meter, the surrounding situation of the device A1, and the like are grasped.

Furthermore, input cells 2-2 and 2-3 are arranged in the device A2. This shows how one device A2 is imaged by two cameras using the cameras CAM-2 and infrared camera CAM-3 of the two input cells 2-2 and 2-3. That is, in device A1, one input cell 2-1 is arranged, but in one device A2, a plurality of input cells are arranged.
Here, the input cell 2-5 images the target area using an infrared camera CAM-3. As a result, the input cell 2-5 can acquire an image or the like that can identify an event such as abnormal heat generation occurring inside the device A2.

Then, each input cell 2-1 to 2-3 converts the data measured by the sensor into linguistic data and sends it to the edge server EDS. In this way, data of each of the plurality of devices A1 and A2 within the premises is collected.

Furthermore, by processing without using control instructions from the central AI 61, it is possible to take action even faster.
FIG. 11 is a diagram illustrating an example of the flow of information processing when performing faster response in the premises of FIG. 10.
In the example shown in FIG. 11, language data is provided from input cell 2-r to output cell 3-q. That is, the language data from the input cell 2-r may be provided to the output cell 3-q without going through the central AI 61.
It is preferable that the input cell 2-r, output cell 3-q, etc. have the following basic logic, especially in the "autonomous decentralized AI blockchain audiovisual cell".

First, it is preferable that the input cell 2-r has "autonomous abnormality detection logic".
That is, the "autonomous abnormality detection logic" is a logic that autonomously analyzes camera images for each input cell 2-r and performs processing to prompt an alarm.
As described above, the input cell 2-r can detect as an abnormal (unusual) state when the image includes an image of a man who does not normally appear. For example, input cell 2-r detects overheating of the device, fire, abnormal behavior (ramping, asking for help, being chased, etc.), intrusion into prohibited areas, intrusion of dangerous animals, etc. It's good to do that.
In addition, as described above with reference to Fig. 5, in addition to a normal camera capable of capturing visible light as shown in Fig. 1, an infrared camera can be used to prevent abnormal overheating inside and outside the equipment, and to detect problems from the other side of the wall. It is also possible to detect phenomena that cannot be detected with normal cameras, such as high-temperature fires that cannot be seen.
Further, it is preferable that the input cell 2-r further includes a microphone as an input section. This makes it possible to use, for example, human voices and sounds generated by accidents, etc. together with images, thereby making it possible to improve detection accuracy.

Second, it is preferable that the input cell 2-r has "past comparison learning (self-learning) logic".
In other words, the "past comparison learning (self-learning) logic" refers to the function of the relearning execution unit 73 of the input cell 2-r, based on information obtained in the past in the input cell 2-r itself and other input cells. This is the logic that learns to perform detection.
Specifically, for example, the relearning execution unit 73 receives an image of a predetermined target person and information regarding the characteristics of the target person from a higher-level server (for example, the edge server EDS in FIG. 5), and displays the image of the target person. re-learning (self-learning) to detect that is actually included in the image. Note that even when the input cell 2-r is detected during relearning (self-learning), the output cell 3-q can be operated to sound an alarm as shown in FIG.

Thirdly, the system preferably includes "auto-tracking logic".
The automatic tracking logic is a logic that automatically tracks a predetermined target in each of the plurality of input cells 2-r using information indicating that the predetermined target is detected in each input cell.
Specifically, for example, a higher-level server (for example, the edge server EDS in FIG. 5) collects information on the detection of a target person or the like by each of the plurality of input cells 2-r. Based on the information that the target person, etc. has been detected at each position in the predetermined area imaged by each input cell 2-r, the higher-level server determines the movement route of the target person, and the future direction of the target person. Automatic tracking by estimating the location.

Fourthly, it is preferable that this system has a "consciousness/intent communication logic."
The consciousness/intent transmission logic is a logic that simultaneously outputs human words such as consciousness and intention in addition to digital image data in each of the plurality of input cells 2-r.
That is, as shown in FIG. 1, the input cell 2-r manages the data of captured images (including still images and moving images) using the blockchain network BCN, and also manages the data of captured images (including still images and video images) using human words. It simultaneously outputs linguistic expressions corresponding to the consciousness and intention of the user.
Specifically, for example, a verbal expression such as "abnormal temperature was detected" or "a person moving violently was detected" is output from the input cell 2-r. By outputting in language expression in this way, as mentioned above, it becomes possible to directly connect to a central AI that can accept existing language input and have it think and make decisions.
Here, the difference between the edge server EDS in FIGS. 5, 6, etc. and the central AI 61 will be explained. That is, as described above, the central AI 61 normally makes basic judgments. However, the edge server EDS can operate the output cell 3-q via the edge server EDS when there is no need to wait for the judgment by the central AI 61.

Fifth, the system preferably has a "self-handling logic".
The self-handling logic is logic in which an operation is executed in each of the plurality of input cells 3-q without going through the central AI 61 or the edge server EDS.
Specifically, unlike a simple surveillance camera, etc., the output cell 3-q outputs an alarm if there is a danger unless immediate action is taken without seeking the judgment of the central AI 61, as explained using FIG. Fires can be extinguished locally using sprinklers, etc.
That is, this realizes information processing when taking high-speed countermeasures.

In this way, the input cell 2-r and output cell 3-q of this information processing system are equipped with audiovisual devices such as cameras and microphones, and realize the concept of identity with IT. It has become something that can be called a cell.

Although one embodiment of the present invention has been described above, the present invention is not limited to the above-described embodiment, and modifications, improvements, etc. within the range that can achieve the purpose of the present invention are included in the present invention. regarded as.

In the above-described embodiment, the input cell 2-r generates language data from data of an image captured by a camera, but the present invention is not limited thereto.
That is, for example, the input cell 2-r may have a function of switching from a "language" for humans to a "language" for machines. Specifically, for example, the input cell 2-r acquires the message "There is a suspicious person" from a human in human language, and indicates "There is a suspicious person" which has been associated in advance in the information processing device. It may also be something that converts into an identifier.
This makes it possible for this information processing system to employ the results of checks made by the human eye as input language data.

As described above, the "autonomous decentralized AI blockchain audiovisual cell" is a unit of information processing device that can execute blockchain-related processing. Note that although the information processing device is used, the present invention is not limited to a personal computer or a server device, and can be implemented as a chip that is an electronic component. This makes it possible to easily incorporate it into various conventional devices.
That is, the autonomous decentralized chip CP in FIG. 1 may be realized as a chip and operated as an input cell by being built into a camera.

These "autonomous decentralized AI blockchain cells" and "autonomous decentralized AI blockchain audiovisual cells" perform autonomous decentralized processing by being connected to each other via a predetermined network, especially a wireless network, within the premises. This can greatly improve convenience.

Regardless of whether wired connection, wireless connection, or a combination thereof is adopted as the network connection method, a large number of input cells 2 ("autonomous decentralized AI blockchain cell" and "autonomous decentralized AI blockchain audiovisual cell") In order to manage, integrate, and utilize them profitably, it is necessary to demonstrate the functions for collaboration.

That is, in a conventional system in which a large number of sensors are managed by a central management server, the load on the central management server increases as the number of sensors increases. However, as mentioned above, by using the input cell 2 of this embodiment, the benefits of decentralized AI are demonstrated as the number of sensors increases, and the central management server (central AI in this embodiment) This will reduce the load. This is also the greatest merit of the input cell 2 of this embodiment.
That is, the most advantageous point of the input cell 2 of this embodiment is that a decentralized system can be easily constructed instead of the conventional centralized system.
The input cell 2 and output cell 3-q including the above-mentioned "autonomous decentralized AI blockchain cell" and "autonomous decentralized AI blockchain audiovisual cell" preferably have the following basic logic.

First, the information processing system of this embodiment preferably has "self-position recognition logic."
That is, in order for the input cell 2 to autonomously determine a solution based on abnormality detection and take action, it is important for the input cell 2 to understand ``where it is'' by itself.
Specifically, for example, when a fire breaks out on the third floor, the input cell 2-1 on the third floor and the input cell 2-2 on the first floor naturally have different priority processes. In other words, distributed processing is required depending on each location and position.

Therefore, for the input cell 2, information such as which device or location on the map the input cell 2 itself is installed in, which input cell 2 among other input cells 2 is nearby, etc. can be determined through the entire map, that is, It is preferable to be able to grasp the relationship between the overall image and one's own position. This enables autonomous control (for example, control processing from the output cell 3) by appropriately utilizing the information of the plurality of input cells 2. .
When mapping this entire map, it is preferable that each input cell 2 is equipped with an information acquisition means for specifying a three-dimensional position, such as a GPS or an altitude sensor. As a result, the location (position) of the input cell 2 itself in the three-dimensional map and the location and device of the abnormality detection destination are not only notified by the higher-level server (for example, the edge server EDS or the server 1) but also by the input cell 2 itself. Can be identified and understood.
Maps are preferably created, updated, and managed by an upper-level server. By periodically notifying each "autonomous decentralized AI blockchain cell", distributed processing can be performed accordingly with a unified map.
Note that by performing this 3D mapping, for example, input cells 2 attached to moving objects such as flying drones can also share mutual positional information. This makes it possible to utilize the input cell 2 to realize safe navigation, proxy navigation, and the like.

Second, it is preferable that the information processing system of this embodiment has an "input cell education logic."
In other words, a higher-level server (for example, edge server EDS or server 1) located inside or outside the campus that manages input cell 2 uses a 3D map of the equipment or department (for an apartment, the room and installation location) and the position and altitude from each input cell. In order to comprehensively map each installed location in 3D based on three-dimensional coordinate information and share it with all input cells 2, it is preferable to periodically notify the latest map.
At this time, in order to always have the latest map, the 3D map can be updated based on information such as whether the input cell 2 is operating normally, movement of location, status, etc. by periodically polling the input cell 2. .
As a result, for example, when a faulty input cell 2 is replaced or a new input cell 2 is added, the newly installed input cell 2 can be automatically input without the need for a human to configure it. It becomes possible for Cell 2 to understand its own position and role and operate accordingly.

Thirdly, the information processing system of this embodiment preferably includes a "distributed autonomous system."
Here, the decentralized autonomous system can be referred to as a "Decentralized Autonomous System (DAS)," and specifically, is as follows.
That is, as described above, the input cell 2 has logic that allows it to autonomously (independently) perform judgment and processing.
Therefore, even if, for example, an on-premises server (for example, edge server EDS) temporarily stops functioning due to a failure or maintenance inspection, the input cell 2 and output cell 3 can be monitored individually, or the input cell 2 and output cell 3 can cooperate with each other. You will be able to take action.
When the on-premises server (for example, edge server EDS) returns (recovers normally), input cell 2 and output cell 3 are updated from the information distributed and held in the upper server (for example, edge server EDS). It is possible to instantly restore the mapping and status of the system and return to normal operation.
The greatest benefit of these input cells 2 and output cells 3 is that they allow the mission to be carried out without any trouble even without a host server (for example, an edge server EDS), and are distributed autonomously. This can be said to be the biggest advantage of the system (DAS).

Fourthly, the information processing system of this embodiment preferably has an "ethical logic."
In other words, the higher-level servers inside and outside the campus (for example, the edge server EDS and server 1) make total predictions and judgments while providing the optimal training for each input cell 2 and output cell 3 based on their own analysis and predictions. conduct.
The term "education" here refers to the fact that the matters to be recognized vary depending on the equipment and environment in which each input cell 2 and output cell 3 is installed.
Specifically, for example, the same sensor information of 30 degrees has different meanings depending on the installed device, such as whether the input cell 2 is installed indoors or outdoors.
Furthermore, for example, noise, vibration, etc. have different meanings depending on the environment and machine.
Therefore, it is preferable that the perceptual expression output by the input cell 2 is corrected so as to produce a verbal output as close as possible to the physical sensation as if a human were sensing it on the spot.
Furthermore, since there are unique words depending on the industry and equipment, it is a good idea to educate students while correcting such unique phrases.
Input cell 2 and output cell 3 receive correction information from these higher-level servers (for example, edge server EDS and server 1), and adjust the output language by themselves while being aware of the location, environment, and equipment in which they are installed. It is preferable to be able to learn.

Furthermore, for example, the system configuration shown in FIG. 2 and the hardware configuration of the server 1 shown in FIG. 3 are merely examples for achieving the object of the present invention, and are not particularly limited.

Further, the functional block diagrams shown in FIGS. 4 and 9 are merely examples, and are not particularly limited. In other words, it is sufficient that the information processing system shown in FIG. 2 is equipped with a function that can execute the above-mentioned series of processes as a whole, and what kind of functional blocks and databases are used to realize this function is particularly dependent on FIG. 4. and is not limited to the example of FIG.

Furthermore, the locations of the functional blocks are not limited to those shown in FIG. 5, and may be arbitrary.
For example, at least a portion of the functional blocks arranged on the server 1 side may be provided in another information processing device.

Furthermore, the series of processes described above can be executed by hardware or by software.
Further, one functional block may be configured by a single piece of hardware, a single piece of software, or a combination thereof.

When a series of processes is executed by software, a program constituting the software is installed on a computer or the like from a network or a recording medium.
The computer may be a computer built into dedicated hardware.
Further, the computer may be a computer that can execute various functions by installing various programs, such as a server, a general-purpose smartphone, or a personal computer.

Recording media containing such programs not only consist of removable media (not shown) that is distributed separately from the main body of the device in order to provide the program to the user, but also are provided to the user in a state that is pre-installed in the main body of the device. Consists of provided recording media, etc.

Note that in this specification, the step of writing a program to be recorded on a recording medium is not only a process that is performed chronologically in accordance with the order, but also a process that is not necessarily performed chronologically but in parallel or individually. It also includes the processing to be executed.

To summarize the above, an information processing system to which the present invention is applied only needs to have the following configuration, and can take various embodiments.

That is, the information processing system to which the present invention is applied (for example, the information processing system of FIG. 4 or FIG. 9),
A central device that executes a predetermined process using input data (for example, the server 1 in FIG. 4 or FIG. 9, or the edge server EDS in FIG. 6 or FIG. 10), and at least a part of the input data to the central device. In an information processing system including one or more type 1 peripheral devices provided (for example, input cell 2 in FIG. 4 or FIG. 9),
The central device includes:
One or more perceptual expressions that acquire one or more perceptual expression data (input language data in FIGS. 4 and 9) as input data, execute a predetermined process using the input data, and show the execution result of the process. Process execution means (process execution unit 51) that outputs data as output data;
Equipped with
Each of the one or more type 1 peripheral devices is
Predetermined data (for example, temperature data or image data) is input, converted into the perceptual expression data (for example, data corresponding to "temperature rise" or data corresponding to "suspicious person is present") and output. model management means (model management unit 71) that stores and manages a model (for example, language conversion AI model 86) in a predetermined storage medium;
Data (e.g., temperature data or image data) output from a sensor that measures a physical quantity in the real world (e.g., sensor 74 in FIG. 4) or a camera that images a target area (e.g., camera in FIG. 9) The perceptual expression data (for example, data corresponding to "temperature rise" or "suspicious person is present") output from the model is input to at least one of the input data of the central device. a conversion means (language conversion AI72) that outputs the language as a part;
It is sufficient to have the following.
This makes it possible to improve the convenience of managing equipment that is managed based on data from a plurality of sensors or cameras.

The information processing system further includes one or more second type peripheral devices (output cells 3) that control a predetermined control target,
Each of the one or more second type peripheral devices is
a model management unit (model management unit 81) that stores and manages a model that inputs the perceptual expression data, converts it into a predetermined physical quantity, and outputs it in a predetermined storage medium;
Acquire at least a part of the one or more perceptual expression data constituting the output data output from the central device and input it to the model, and send a signal indicating the predetermined physical quantity output from the model to an instruction signal. A control means (inverse language conversion AI 82 and control unit 83) that controls the predetermined control object by inputting it to the predetermined control object as
can be provided.
This can further improve the convenience of managing equipment that is managed based on data from a plurality of sensors or cameras.

The central device includes:
Position information from the one or more Type 1 peripheral devices (for example, three-dimensional position information that can be obtained from a GPS or an altitude sensor) and information about the model (for example, the model itself, which affects a threshold value for determining a temperature increase) data, etc.) as first type peripheral device information;
map management means (for example, processing execution unit 51) that generates or updates and manages a three-dimensional map of equipment in which the one or more first type peripheral devices are arranged, based on the first type peripheral device information;
further comprising;
Each of the one or more first type peripheral devices is
a position information acquisition unit (for example, model management unit 71) that acquires position information of the first type peripheral device;
a first type peripheral device information transmitting means (for example, model management section 71) that provides the central device with information regarding the location information and the model;
a first type relearning execution unit (for example, the relearning execution unit 73 in FIG. 4) that executes relearning for updating the model based on the three-dimensional map managed in the central device and the position information; ,
It is possible to further include the following.
This can further improve the convenience of managing equipment that is managed based on data from a plurality of sensors or cameras.

DESCRIPTION OF SYMBOLS 11... Server, 2... Input cell, 3... Output cell, 4... Confirmation terminal, 5... Blockchain full node, 11... CPU, 20... Drive, 31... ...Removable media, 51...Process execution unit, 61...Central AI, 71...Model management unit, 72...Language conversion AI, 73...Relearning execution unit, 74...Sensor , 75... Language conversion AI model, CAM... Camera, 81... Model management unit, 82... Reverse language conversion AI, 83... Control unit, 84... Relearning execution unit, 85 ...actuator, 86...reverse language conversion AI model

Claims

An information processing system including a central device that executes predetermined processing using input data, and one or more type 1 peripheral devices that provide at least a part of the input data to the central device,
The central device includes:
processing execution means that acquires one or more perceptual expression data as input data, executes a predetermined process using the input data, and outputs one or more perceptual expression data indicating the execution result of the process as output data;
Equipped with
Each of the one or more type 1 peripheral devices is
a model management means for storing and managing a model that inputs predetermined data and converts it into the perceptual expression data and outputs it in a predetermined storage medium;
Data output from a sensor that measures physical quantities in the real world or a camera that images a target area is acquired and input to the model, and the perceptual expression data output from the model is input to the central device. a conversion means outputting as at least part of the data;
Equipped with
Information processing system.
The information processing system further includes one or more second type peripheral devices that control a predetermined control target,
Each of the one or more second type peripheral devices is
a model management means for storing and managing a model that inputs the perceptual expression data, converts it into a predetermined physical quantity, and outputs it in a predetermined storage medium;
Acquire at least a part of the one or more perceptual expression data constituting the output data output from the central device and input it to the model, and send a signal indicating the predetermined physical quantity output from the model to an instruction signal. a control means for controlling the predetermined control object by inputting it to the predetermined control object;
The information processing system according to claim 1, comprising:
The central device includes:
first type peripheral device information acquisition means for respectively acquiring location information and information regarding the model from the one or more first type peripheral devices as first type peripheral device information;
map management means for generating or updating and managing a three-dimensional map of equipment in which the one or more first type peripheral devices are arranged based on the first type peripheral device information;
further comprising;
Each of the one or more first type peripheral devices is
a position information acquisition means for acquiring position information of the first type peripheral device;
first type peripheral device information transmitting means for providing the central device with information regarding the location information and the model;
a first type relearning execution means for executing relearning for updating the model based on a three-dimensional map managed in the central device and the position information;
further comprising,
The information processing system according to claim 2.
An information processing method executed by an information processing system including a central device that executes a predetermined process using input data, and one or more first type peripheral devices that provide at least a part of the input data to the central device,
The steps performed by the central device include:
a processing execution step of acquiring one or more perceptual expression data as input data, executing a predetermined process using the input data, and outputting one or more perceptual expression data indicating the execution result of the process as output data;
including,
The steps performed by each of the one or more first type peripheral devices include:
a model management step of inputting data of a predetermined image and storing and managing a model to be converted into the perceptual expression data and outputted in a predetermined storage medium;
Data output from a sensor that measures physical quantities in the real world or a camera that images a target area is acquired and input to the model, and the perceptual expression data output from the model is input to the central device. a transformation step outputting as at least a portion of the data;
including,
Information processing method.
A computer that controls an information processing system that includes a central device that executes predetermined processing using input data, and one or more first type peripheral devices that provide at least a part of the input data to the central device,
a computer controlling the central device;
a processing execution step of acquiring one or more perceptual expression data as input data, executing a predetermined process using the input data, and outputting one or more perceptual expression data indicating the execution result of the process as output data;
Execute control processing including
a computer that controls each of the one or more type 1 peripheral devices;
a model management step of inputting data of a predetermined image and storing and managing a model to be converted into the perceptual expression data and outputted in a predetermined storage medium;
Data output from a sensor that measures physical quantities in the real world or a camera that images a target area is acquired and input to the model, and the perceptual expression data output from the model is input to the central device. a transformation step outputting as at least a portion of the data;
execute control processing including
program.