WO2019015615A1

WO2019015615A1 - Fog node deployment method and fog network system

Info

Publication number: WO2019015615A1
Application number: PCT/CN2018/096165
Authority: WO
Inventors: 刘芙蕾
Original assignee: 中兴通讯股份有限公司
Priority date: 2017-07-19
Filing date: 2018-07-18
Publication date: 2019-01-24
Also published as: CN109286508A

Abstract

A fog node deployment method and a fog network system. The fog node deployment method comprises: providing at least one layer of fog nodes to form a fog network, and configuring the fog network to obtain data from a collection device, perform machine learning according to the data to obtain data law information, and control an execution device according to the data law information.

Description

Fog node deployment method and fog network system

Technical field

The present disclosure relates to, but is not limited to, the field of edge computing.

Background technique

The Open Fog Alliance (OpenFog) aims to accelerate the deployment of fog technology by developing core technologies such as open architecture, distributed computing, networking and storage, and the leadership needed to realize the full potential of the Internet of Things. OpenFog's mission is to drive industrial and academic research in fog computing architectures, test development, interoperability, and assembly lines to seamlessly connect edge-to-cloud architectures, enabling end-to-end Internet of Things (IOT) scenarios. Become a reality. OpenFog's reference architecture is a vertical, system-level architecture that distributes compute, storage, communication, control, and network functions closer to users. Its reference architecture represents a traditional closed system and a cloud-only deployment model. The shift, which shifts to a new computing model, is about moving from the cloud to the edge, even on IoT sensors and actuators. The computation, network, storage, and acceleration units of the new model can all be fog nodes. Each layer in the layered architecture consisting of fog nodes provides additional processing, storage, and networking capabilities for vertical application at that layer.

The fog node proposed by OpenFog has complete functions from hardware to software. In the reference architecture released by it, the application scenarios of various fog nodes are given, but there is no explanation on how to deploy the fog nodes and what kind of fog nodes are implemented in different magnitudes. Features.

Public content

Embodiments of the present disclosure provide a fog node deployment method, including: providing at least one layer of fog nodes to form a fog network; and configuring the fog network to acquire data from a collection device and perform machine learning based on the data To obtain data regularity information, and then control the execution type device according to the data regularity information.

Embodiments of the present disclosure provide a storage medium storing a program including executable instructions that are executed to implement the above-described fog node deployment method.

An embodiment of the present disclosure provides a fog network system, including: at least one layer of fog nodes configured to acquire data from a collection device, and perform machine learning according to the data to obtain data regularity information, and then according to the data rule Information to control the execution class device.

DRAWINGS

FIG. 1 is a flowchart of a method for deploying a fog node according to an embodiment of the present disclosure;

2 is a schematic diagram of a single fog node deployment manner provided by an embodiment of the present disclosure;

3 is a schematic diagram of a multi-layer fog node deployment manner provided by an embodiment of the present disclosure;

4 is a data reporting flowchart of a single fog node provided by an embodiment of the present disclosure;

5 is a flow chart of issuing a command of a single fog node according to an embodiment of the present disclosure;

FIG. 6 is a flow chart of data reporting of a multi-layered fog node provided by an embodiment of the present disclosure.

Detailed ways

The embodiments of the present disclosure are described in detail below with reference to the accompanying drawings.

The embodiments of the present disclosure are applicable to scenarios such as smart cities, for example, for managing and controlling the power usage of a building, a cell, or a city in a city or a city. For specific applications, the fog network performs machine learning based on power data collected from building/community/region/city-wide collection devices, and controls building/community/region/city-wide execution-type devices based on machine learning results.

FIG. 1 is a flowchart of a fog node deployment method provided by an embodiment of the present disclosure. As shown in FIG. 1, the fog node deployment method includes the following steps S101 and S102.

Step S101: Provide at least one layer of fog nodes to form a fog network.

Step S102: Configure the fog network to acquire data from the collection device, perform machine learning according to the data, obtain data regularity information, and then control the execution device according to the data regularity information.

In some embodiments, the fog network may include only one layer of fog nodes, ie, edge layer fog nodes, which may connect devices through the access network, and collect data collection, processing, machine learning, communication, devices. The control and other functions are integrated, and it is a centralized fog node.

For a fog network comprising only one layer of fog nodes, the fog node (ie, the edge layer fog node) may be configured to acquire data from the collection class device and perform machine learning based on the data to obtain data regularity information (ie, Data regularity information of the edge layer). The fog node may be further configured to send data rule information of the edge layer to the cloud platform.

In some embodiments, the mist network can include two layers of fog nodes, namely an edge layer fog node and a high level fog node. The edge layer fog node can connect to the device through the access network, which can have only data collection, pre-processing and communication functions, and can also have functions such as machine learning and device control. The high-level fog node can be connected to the edge layer fog node, and can have functions such as data processing, machine learning, communication, and device control, and can also have a fog node management function.

For a fog network including two layers of fog nodes, the edge layer fog node acquires data from the collection class device and transmits the data to the high-level fog node, or the edge layer fog node performs machine learning according to the data acquired from the collection class device, The data regularity information of the edge layer is obtained and sent to the high-level fog node. The high-level fog node performs machine learning according to the data sent by the edge layer fog node to obtain high-level data regularity information. The high-level fog node may also send the high-level data rule information to the cloud platform.

In some embodiments, the mist network can include three layers of fog nodes, namely an edge layer fog node, a middle layer fog node, and a high level fog node. The edge layer fog node can connect to the device through the access network, which can have only data collection, pre-processing and communication functions, and can also have functions such as machine learning and device control. The middle fog node can be connected to the edge layer fog node and can have functions such as data collection, processing, machine learning and device control. The high-level fog node can be connected to the middle fog node, and can have functions such as data processing, machine learning, communication, and device control, and can also have fog node management functions.

For the fog network including the three-layer fog node, the edge layer fog node acquires data from the collection class device and transmits the data to the middle layer fog node, or the edge layer fog node performs the machine according to the data acquired from the collection class device. Learning to obtain the data regularity information of the edge layer and send it to the middle layer fog node. The middle fog node performs machine learning according to the data sent by the edge layer fog node to obtain the data regularity information of the middle layer, and sends the data to the high-level fog node. The high-level fog node performs machine learning according to the data regularity information of the middle layer to obtain high-level data regularity information. The high-level fog node may also send the high-level data rule information to the cloud platform.

In some embodiments, the mist network includes a plurality of layers of fog nodes, specifically, based on a fog network formed by edge layer fog nodes, middle layer fog nodes, and high-level fog nodes, edge layer fog nodes, middle layer fog nodes At least one of the high-rise fog nodes may include two or more layers of fog nodes to form a fog network that actually includes more than three layers of fog nodes. In other words, each of the three-layered fog nodes may have two or more layers of fog nodes, thereby obtaining a fog network including three or more fog nodes.

For the fog network including only one layer of fog nodes, the fog node (ie, the edge layer fog node) may generate a control command (ie, an edge according to the data rule information (ie, data rule information of the edge layer) Layer control commands) and send them to the execution class device to implement edge layer control of the execution class device. The cloud node (ie, the edge layer fog node) may further transmit the cloud control command generated and delivered by the cloud platform according to the data rule information of the edge layer to the execution device to implement the cloud of the execution device. control.

For the fog network including two layers of fog nodes, the high-level fog node may generate a high-level control command according to the data rule information of the upper layer, and send the same to the execution device through the edge layer fog node to implement the execution. High-level control of class devices. The edge layer fog node may further generate an edge layer control command according to the data rule information of the edge layer, and send the edge layer control command to the execution class device to implement edge layer control on the execution class device. The high-level fog node can also transmit the cloud control command generated and delivered by the cloud platform according to the data rule information of the upper layer to the execution class device through the edge layer fog node to implement cloud control on the execution class device.

For the fog network including the three-layer fog node, the middle layer fog node generates a middle layer control command according to the data rule information of the middle layer, and sends the middle layer control command to the execution class device through the edge layer fog node to implement the execution class. Middle layer control of the device. The high-level fog node generates a high-level control command according to the high-level data rule information, and sequentially sends the high-level control command to the execution class device through the middle layer fog node and the edge layer fog node to implement high-level control of the execution class device. The edge layer fog node may also generate an edge layer control command according to the data rule information of the edge layer, and send the edge layer control command to the execution class device to implement edge layer control on the execution class device. The high-level fog node may further pass the cloud control command generated and delivered by the cloud platform according to the data rule information of the upper layer to the execution class device through the middle layer fog node and the edge layer fog node, so as to implement the execution class. Cloud control of the device.

The function configuration of each fog node in the fog network including the three or more fog nodes may be referred to the function configuration of each fog node in the fog network including the three-layer fog node, and details are not described herein again.

The collection device and the execution device may be the same device or different devices, that is, the collection device and the execution device may be one physical entity or separate devices.

The above device may be an Internet of Things device, such as a sensor, a physical tag, an electric appliance, or the like, or may be a terminal device such as a mobile phone, a computer, or the like.

It can be understood by those skilled in the art that all or part of the steps of the fog node deployment method of the above embodiment may be completed by hardware related to program instructions, and the program may be stored in a computer readable storage medium, and the program is executed. The fog node deployment method of the above embodiment is implemented.

Embodiments of the present disclosure also provide a storage medium having stored thereon a program including executable instructions that, when executed by, for example, a computer or a processor, implement the fog node deployment method of the above embodiment. The storage medium may include a ROM/RAM, a magnetic disk, an optical disk, a USB flash drive, or the like.

Embodiments of the present disclosure also provide a fog network system, including: at least one layer of fog nodes configured to acquire data from a collection class device and perform machine learning based on the data to obtain data rule information, and then according to the data Regular information controls the execution class.

The fog network may include only one layer of fog nodes, that is, edge layer fog nodes, and the edge layer fog nodes may be configured to acquire data from the collection class device and perform machine learning according to the data to obtain edge layer data. The rule information is then generated according to the data rule information of the edge layer, and is sent to the execution class device to implement edge layer control on the execution class device.

The fog network may include two layers of fog nodes, namely an edge layer fog node and a high layer fog node. The edge layer fog node may acquire data from the collection class device and send the data to the high-level fog node, or perform machine learning according to data acquired from the collection class device to obtain data regularity information of the edge layer. And send it to the high-level fog node. The high-level fog node performs machine learning according to the data sent by the edge layer fog node to obtain high-level data rule information, and then generates a high-level control command according to the high-level data rule information, and uses the edge layer fog node to Sending to the execution class device to implement high level control of the execution class device.

The fog network may include three layers of fog nodes, namely an edge layer fog node, a middle layer fog node, and a high level fog node. The edge layer fog node acquires data from the collection class device and sends the data to the middle layer fog node, or performs machine learning according to the data acquired from the collection class device to obtain data regularity information of the edge layer, and Send to the middle fog node. The middle fog node performs machine learning according to the data sent by the edge layer fog node to obtain the data regularity information of the middle layer, and sends the data to the high-level fog node, and then generates a middle layer control command according to the data rule information of the middle layer, and The edge layer fog node is sent to the execution class device to implement middle layer control of the execution class device. The high-level fog node performs machine learning according to the data regularity information of the middle layer to obtain high-level data rule information, and then generates a high-level control command according to the high-level data rule information, and sequentially passes through the middle layer fog node and the edge. The layer fog node sends it to the execution class device to implement high level control of the execution class device.

The fog network may include three or more fog nodes, that is, at least one of the edge layer fog node, the middle layer fog node, and the high-level fog node may include two or more layers of fog nodes, thereby forming a fog node including three or more layers. Fog network. For the function configuration of each node in the fog network including the three or more fog nodes, refer to the function configuration of each fog node in the fog network including the three-layer fog node, and details are not described herein again.

In some embodiments, the fog network is further configured to send the data rule information to the cloud platform, and transmit the cloud control command generated and delivered by the cloud platform according to the data rule information to the execution class. A device to implement cloud control of the execution class device. Specifically, the fog network may be obtained by deploying a single fog node as shown in FIG. 2, or by deploying two layers of fog nodes, or may be obtained by deploying a three-layer fog node as shown in FIG. 3, and may also be deployed. More than three layers of fog nodes are obtained.

In the same fog network, all the fog nodes are impossible to perform the same function, and do not need to have the same capabilities. According to the application scenario, there are various types and functions of fog nodes. Depending on the application, the fog node can be defined to include an edge/light fog node, a middle fog node, and a high level fog node, which can be deployed to form a fog network as shown in FIG.

The edge/light fog node is only used for data collection and communication. Its main functions include: (1) acquisition and collection of sensor data, such as periodic or event-triggered acquisition of physical device data; (2) processing of data format (3) passing data to the upper fog node or the cloud; (4) controlling the sensor and the actuator, and issuing commands to the upper node to the sensor and the actuator.

The middle fog node is mainly used for data collection, processing and communication, specifically for: (1) data collection, such as collecting data from each edge fog node; (2) data filtering, compression, merging, format conversion, simple data analysis Etc., for example, packet filtering and culling invalid information according to certain rules; (3) communication between north-south fog/cloud nodes and east-west fog nodes: a. uploading and collecting edge fog nodes After the necessary processing, the data and information are transmitted to the upper fog node/cloud, and the upper fog node/cloud control command is transmitted to the edge fog node; b. The necessary data transmission between the left and right fog nodes, information sharing Etc., such as link quality, routing information, and load information; (4) machine learning, analyzing long-term collection of fixed area or user data, machine learning from time, space, etc., to find models or rules Provide users with effective information to help with policy development.

The high-level fog nodes do not collect data. The main functions include: (1) data processing; (2) network management, such as fog node management; (3) big data analysis, machine learning; (4) communication information transfer, such as node and Communication between nodes, communication between nodes and the cloud.

According to the function configuration of the fog node, the fog node can be deployed flexibly according to the actual situation. That is, for the specific application scenario, some two fog nodes can be merged, for example, the edge fog node and the middle fog node are merged, or the middle fog node is merged. Merged with high-level fog nodes.

Typical deployment methods of fog nodes are as follows: (1) deploying a layer of fog nodes to form a fog network; (2) deploying two layers of fog nodes to form a fog network; and (3) deploying three layers of fog nodes to Form a fog network.

Deploying a layer of fog nodes can also be considered as a thin deployment or a centralized deployment. The deployment method is shown in Figure 2. In this deployment mode, sensors and actuators can be connected to the network through wireless means (such as LTE/5G/WIFI, etc.) or wired (such as Ethernet), and centralized fog nodes connect to sensors and execute in the network. Data collection, data analysis, machine learning and control. In this way, the functions of the edge fog node, the middle fog node, and the high-level fog node are combined, that is, the edge fog node performs all data collection, data processing, and data analysis.

This deployment method is applied to small, regional scenarios where the number of sensors and actuators is small, the reported data and the amount of data required are small, and the real-time performance is low. A fog node (ie, an edge fog node) can Complete data storage and processing.

In the deployment mode of the fog network formed by the edge node and the high-level node, the sensor and the actuator are connected to the network through a wireless method (such as LTE/5G/WIFI, etc.) or a wired method (such as Ethernet), and the edge fog node is responsible for completing Raw data collection and simple processing. For data with high real-time requirements and large bandwidth transmission, after the edge fog node completes the data analysis, it sends an execution command to the sensor and the actuator, and then the edge fog node backs up the analysis result periodically or The event is triggered to the high-level fog node, and the high-level fog node only needs to know the processing result and the valid data for big data analysis. In this deployment mode, the edge fog node actually completes the functions of the edge fog node and the middle layer fog node.

The deployment mode is applied to a scenario where the number of sensors and actuators is large and distributed, and the types of access modes (wired or wireless) are high, real-time performance is high, and local data volume is large.

As shown in FIG. 3, the fog network is configured by an edge node, a middle node, and a high-level node. In this mode, the sensors and actuators are connected to the network through wireless means (such as LTE/5G/WIFI, etc.) or wired (such as Ethernet), and the edge fog node is responsible for collecting the original data and processing the data format. If the collected data has high real-time requirements and requires large bandwidth transmission, the edge fog node also needs to complete the local analysis, send the execution command to the sensor and the actuator, and send the analysis result to the middle fog node. The middle fog node performs data aggregation of multiple edge fog nodes, and further analyzes the data, and performs data forwarding in the east-west or north direction. When the edge fog node fails, the middle fog node can directly perform the fault fog node. Alerts and backs up the fog node data. The high-level fog node manages the middle fog node, coordinates the load of each middle fog node, analyzes the data returned by the middle fog node, and gives the intelligent result, which is sent to the middle fog node and the edge fog node for execution.

In the above deployment mode, the deployment mode of the three-layer fog node can be extended to the multi-layer fog node deployment mode. Depending on the requirements of the actual application scenario, it can be between the edge fog node and the middle fog node or between the middle fog node and the high-level fog node. Then add the middle fog node, which is derived into the multi-layer fog node deployment mode. In general, two-layer fog node deployment and three-layer fog node deployment are common and typical deployment methods.

The following is an example of a fog node deployment mode shown in FIG. 2 .

In the scenario of intelligently managing the air conditioning system of a building and deploying a fog node in the building, the fog node can have the following functions: 1. Support IOT equipment and gateway device access, for example, support narrowband Internet of Things ( NB-IOT), Long-Term Evolution (LTE), Long-Range Coverage (Lora) and other access technologies. Take LTE indoor deployment Qcell as an example. Telecom operators deploy Qcell base stations in the building to ensure wireless coverage of the entire building. That is, the access network in Figure 2 is Qcell; 2. Supports data acquisition of IOT devices (sensors and actuators), for example, collecting data in a periodic manner or in an event-triggered manner, and the IOT device for data collection is hand-held by users in the building. The terminal (for example, a mobile phone), the user's mobile phone accesses the Qcell network, and periodically reports information to the Qcell base station, for example, it may be uplink power information, and the format of the report information may be defined by a 3GPP Qcell, and the Qcell base station directly reports the information. Passed to the fog node, the data format of the Qcell split can be IP data packets or UDP data packets; 3. Support the edge calculation of the collected data, that is, data filtering, Analysis, machine learning, to obtain a relevant time dimension or model of the flow dimension, for example, the fog node performs machine learning based on the information reported by the mobile phone, analyzes the distribution of people flow on each floor and each room, and the distribution and time of people flow ( For example, a relationship of 24 hours a day, 7 hours a day, that is, a relationship model of "flow-floor/room-time" is obtained. After obtaining the model, in turn, according to the flow of people, the air conditioning temperature of each floor/room can be dynamically adjusted in different time periods, thereby achieving the purpose of energy saving. Specifically, when the flow rate of the person is large, the air conditioner temperature is lowered, when the flow rate of the person is small, the air conditioner temperature is raised, or the air conditioner is turned off. That is, the fog node controls the accessed IOT device according to the machine learning result, where the IOT device that executes the command is the air conditioner controller of the building; 4. supports the edge application (APP), for example, the edge application may be an air conditioner energy saving, actually Up, the fog node can be deployed for a variety of applications and services, such as building energy (lighting, electricity, etc.), location services, etc.; 5. can be connected to the cloud to achieve cloud, fog combined applications (APP), for example, The fog node deployed in the building can also be connected to the cloud platform. The fog node only needs to transmit the analysis result to the cloud platform for periodic (periodical or event-triggered) or further data analysis on the cloud platform. In addition, the cloud platform You can issue commands or you can directly issue commands from the fog node.

4 is a data reporting flowchart of a single fog node provided by an embodiment of the present disclosure. As shown in FIG. 4, the data reporting of a single fog node includes steps 501 to 506.

Steps 501 and 502: The data generated by the IOT device (sensor or terminal) is reported to the Qcell base station through the 3GPP standard interface.

Steps 503 to 505: The Qcell base station divides the collected data into the fog node in the format of the IP packet, and the fog node completes the data analysis, and the fog node can directly issue the command to the IOT device according to the analysis result, and the fog node can analyze the result. After uploading to the cloud platform, after the cloud platform collects the data, it can perform further analysis and machine learning according to the received data of multiple fog nodes to generate commands.

FIG. 5 is a flowchart for issuing a command of a single fog node according to an embodiment of the present disclosure. As shown in FIG. 5, the command issuing of a single fog node includes steps 601 to 606.

Step 601: The cloud platform sends the generated command to the corresponding fog node.

Steps 602 to 604: The fog node transmits the command of the cloud to the Qcell base station, and when the command is delivered, the fog node performs necessary data format conversion.

Step 605: The Qcell base station transmits a command to the IOT device (executor).

Step 606: The IOT device (executor) receives and executes the command.

The following is an example of the three-layer fog node deployment mode shown in FIG. 3. As shown in FIG. 3, an IOT device (sensor, actuator), an access network, an edge node, a middle node, and a high-level node form a network. Take City A's power system as an example. For example, IOT equipment is a card reader chip on the electricity meter of each building in city A. The access network is the NB-IOT network deployed by city A, and the edge node is the fog node of each building of city A. The middle node is the fog node deployed in each area of the city A, and the high-level node is the urban fog node deployed by the city A.

In this case, the IOT equipment of each building is connected to the network through the NB-IOT network, and each IOT device reads the data of the electricity meter and periodically reports it to the NB-IOT base station, and the NB-IOT base station distributes the data to each building. Edge fog node. The edge fog node completes the following main functions: a) Through machine learning, the analysis of the distribution characteristics of the entire building is completed, and the corresponding characteristic curve of power consumption and time is generated; b) the command can be generated according to the working condition of the IOT device, for example When an IOT device fails, the fog node can alert the maintenance personnel to remind the fault that it needs to be repaired. The fog node can also call the device adjacent to the faulty IOT device to take over the task of the faulty device; c) The curve is uploaded to the middle fog node of the area, and the command of the middle fog node is received. The middle fog node completes the following main functions: a) After receiving the data of each edge fog node, it performs machine learning on the power consumption of the building in the entire area, generates a heat map of the distribution of power usage in the area, and the power consumption and time of the area. The corresponding curve, so as to make recommendations for the power supply in the entire area, and the dynamic adjustment of the power consumption in different time periods under the same power supply; b) manage the edge fog nodes in the area, when there are new edge fog nodes to join When the edge fog node fails, the data can be backed up in time to generate a warning, and the adjacent backup edge fog node can take over the fault edge fog node; c) upload the analysis result. Go to the high-level fog node and receive the command of the high-level fog node. The high-level fog node completes the following main functions: a) Machine learning based on the collected data of the middle fog node, generating a heat map of the electricity consumption of each administrative area in the city, and a corresponding curve of the electricity consumption and time of each district, for the entire city Suggestions on the use of electricity, for example, the central city is the peak of electricity during the day, the low peak of electricity during the night, the low peak of electricity during the daytime in the suburbs, and the peak of electricity consumption at night, when the total power supply of the whole city is constant. According to the power consumption during the day and night, the power supply of each area can be dynamically adjusted to save energy; b) the command is given to the middle fog node, for example, how much power is used in a certain period of time; c) the data can be uploaded to The cloud platform is capable of executing and releasing commands from the cloud platform.

FIG. 6 is a flow chart of data reporting of a multi-layered fog node provided by an embodiment of the present disclosure. As shown in FIG. 6, the data reporting of the multi-layer fog node includes steps 701 to 710.

Steps 701 to 702: The data generated by the IOT device (sensor or terminal) is reported to the NB-IOT base station through the 3GPP standard interface.

Steps 703 to 705: The NB-IOT base station divides the collected data into an edge fog node in an IP packet format, and the edge fog node performs data extraction and analysis.

Step 706: The edge fog node reports the processed data to the middle layer fog node.

Steps 707 to 708: The middle layer fog node performs data analysis, and uploads the analysis result to the high-level fog node.

Steps 709 to 710: The high-level fog node performs machine learning according to the collected data of the middle-level fog node, and uploads the analysis result to the cloud platform.

In addition, the fog nodes at each level can generate commands themselves, or receive commands from the upper node, and then send them to the sensor/actuator through the next level node.

Embodiments of the present disclosure can realize the object/person-edge calculation-cloud computing in the north-south direction by functionally defining and deploying the fog node, and can realize the east-west person and person, the person and the object, the object and the object. Free and dynamic connectivity and communication, as well as the ability to dynamically share computing resources, storage resources, etc., accelerate local computing speed, and improve efficiency.

While the embodiments of the present disclosure have been described in detail above, the present disclosure is not limited thereto, and various modifications and changes can be made by those skilled in the art in accordance with the principles of the present disclosure. Modifications and variations of the present invention are intended to be included within the scope of the present disclosure.

Claims

A method for deploying a fog node, comprising:

Providing at least one layer of fog nodes to form a fog network;

The fog network is configured to acquire data from the collection device, perform machine learning according to the data, obtain data regularity information, and then control the execution device according to the data regularity information.
The method according to claim 1, wherein the fog network comprises an edge layer fog node, and the edge layer fog node acquires and machine learning data of the collection class device to obtain data regularity information of the edge layer, and then And generating an edge layer control command according to the data rule information of the edge layer, and sending the edge layer control command to the execution class device to implement edge layer control on the execution class device.
The method of claim 1 wherein said mist network comprises an edge layer fog node and a high level fog node, said edge layer fog node acquiring data from said collection class device and obtaining from said collection class device The data is sent to the high-level fog node, and the high-level fog node performs machine learning according to the data sent by the edge layer fog node to obtain high-level data rule information, and generates a high-level control command according to the high-level data rule information. The high layer control command is then sent to the execution class device by the edge layer fog node to implement high level control of the execution class device.
The method of claim 1, wherein the fog network comprises an edge layer fog node and a high level fog node, the edge layer fog node acquiring data from the collection class device and according to the acquisition from the collection class device The data is machine-learned, the data regularity information of the edge layer is obtained, and the data regularity information of the edge layer is sent to the high-level fog node, and the high-level fog node performs machine learning on the data sent by the edge layer fog node. Obtaining high-level data rule information, and generating a high-level control command according to the high-level data rule information, and then sending the high-level control command to the execution class device through the edge layer fog node to implement the execution High-level control of class devices.
The method according to claim 4, wherein the edge layer fog node further generates an edge layer control command according to the data rule information of the edge layer, and sends the edge layer control command to the execution class device. To implement edge layer control of the execution class device.
The method of claim 1, wherein the mist network comprises an edge layer fog node, a middle layer fog node, and a high level fog node, the edge layer fog node acquiring data from the collection class device and collecting from the collection The data acquired by the class device is sent to the middle fog node, and the middle fog node performs machine learning according to the data sent by the edge layer fog node to obtain data regularity information of the middle layer, and sends the data regularity information of the middle layer to the high layer fog. a node, the middle layer fog node further generates a middle layer control command according to the data rule information of the middle layer, and sends the middle layer control command to the execution class device by using the edge layer fog node to implement the performing The middle layer control of the class device, the high-level fog node performs machine learning on the data regularity information of the middle layer, obtains high-level data rule information, and generates a high-level control command according to the high-level data rule information, and then sequentially passes the The middle fog node and the edge layer fog node send the high layer control command to the execution class device to Implementing high level control of the execution class device.
The method of claim 1, wherein the mist network comprises an edge layer fog node, a middle layer fog node, and a high level fog node, the edge layer fog node acquiring data from the collection class device, according to the collection class The data acquired by the device is machine-learned to obtain data rule information of the edge layer, and the data rule information of the edge layer is sent to the middle layer fog node, and the middle layer fog node performs data according to the data sent by the edge layer fog node. The machine learning obtains the data regularity information of the middle layer, and sends the data regularity information of the middle layer to the high-level fog node, and the middle layer fog node further generates a middle layer control command according to the data rule information of the middle layer, and passes the edge The layer fog node sends the middle layer control command to the execution class device to implement middle layer control on the execution class device, and the high layer fog node performs machine learning according to the middle layer data rule information to obtain high layer data. Regular information, and generate high-level control commands according to the high-level data rule information, and then sequentially Said intermediate layer and said edge node mist mist of the high level node transmits a control command to perform the type of equipment, in order to achieve high-level control of the execution class device.
The method according to claim 7, wherein the edge layer fog node further generates an edge layer control command according to the data rule information of the edge layer, and sends the edge layer control command to the execution class device. To implement edge layer control of the execution class device.
The method of claim 6, wherein at least one of the edge layer fog node, the middle layer fog node, and the high level fog node comprises two or more layers of fog nodes.
The method of claim 7, wherein at least one of the edge layer fog node, the middle layer fog node, and the high level fog node comprises two or more layers of fog nodes.
A fog network system comprising:

At least one layer of fog nodes is configured to acquire data from the collection device, perform machine learning according to the data, obtain data rule information, and then control the execution device according to the data rule information.
The system of claim 11, wherein the at least one layer of fog nodes comprises an edge layer fog node configured to acquire data from the collection class device and perform machine learning to obtain data rule information of the edge layer, and then according to The data layer information of the edge layer generates an edge layer control command, and sends the edge layer control command to the execution class device to implement edge layer control on the execution class device.
The system of claim 11 wherein said at least one layer of fog nodes comprises an edge layer fog node and a high level fog node, said edge layer fog node configured to acquire data from said collection class device and to The data obtained by the collection device is sent to the high-level fog node, and the high-level fog node is configured to perform machine learning according to the data sent by the edge layer fog node, obtain high-level data regularity information, and according to the high-level data rule The information is generated, and a high-level control command is generated, and then the high-level control command is sent to the execution class device by the edge layer fog node to implement high-level control of the execution class device.
The system of claim 11 wherein said at least one layer of fog nodes comprises an edge layer fog node and a high level fog node, said edge layer fog node configured to acquire data from said collection class device and according to said Collecting data acquired by the class device for machine learning, obtaining data rule information of the edge layer, and transmitting data rule information of the edge layer to the high-level fog node, where the high-level fog node is configured according to the edge layer fog node Transmitting the data for machine learning, obtaining high-level data rule information, and generating a high-level control command according to the high-level data rule information, and then sending the high-level control command to the execution device through the edge layer fog node To achieve high-level control of the execution class device.
The system according to claim 14, wherein the edge layer fog node further generates an edge layer control command according to the data rule information of the edge layer, and sends the edge layer control command to the execution class device. To implement edge layer control of the execution class device.
The system of claim 11 wherein said at least one layer of fog nodes comprises an edge layer fog node, a middle layer fog node, and a high level fog node, said edge layer fog node configured to acquire data from said collection class device, and Transmitting data acquired from the collection device to a middle fog node, wherein the middle fog node is configured to perform machine learning according to data sent by the edge layer fog node, to obtain data regularity information of the middle layer, and to obtain the middle layer data The data regularity information is sent to the high-level fog node, and the middle-layer fog node is further configured to generate a middle-level control command according to the data rule information of the middle layer, and then send the middle-level control command to the execution by using the edge layer fog node And the high-level fog node is configured to perform machine learning according to the data regularity information of the middle layer, obtain high-level data regularity information, and according to the high-level data rule information, the class device is configured to implement the middle-level control of the execution class device. Generating a high-level control command, and then sequentially passing through the middle layer fog node and the edge layer fog node Layer control execution command is transmitted to the device type, in order to achieve high-level control of the execution class device.
The system of claim 11 wherein said at least one layer of fog nodes comprises an edge layer fog node, a middle layer fog node, and a high level fog node, said edge layer fog node configured to acquire data from said collection class device, and Performing machine learning according to the data acquired by the collection device, obtaining data rule information of the edge layer, and transmitting data rule information of the edge layer to the middle layer fog node, where the middle layer fog node is configured according to the edge The data sent by the layer fog node performs machine learning to obtain data regularity information of the middle layer, and sends the data regularity information of the middle layer to the high-level fog node, and the middle layer fog node is further configured to generate according to the data rule information of the middle layer. a middle layer control command, and then sending, by the edge layer fog node, the middle layer control command to the execution class device to implement middle layer control on the execution class device, where the high layer fog node is configured according to the middle layer Data regularity information is machine learning, obtaining high-level data law information, and according to the high-level data law Information generating layer control command, followed by the middle layer and said edge node mist mist of the high level node is transmitted to the control command execution by the device type, in order to achieve high-level control of the execution class device.
The system according to claim 17, wherein the edge layer fog node further generates an edge layer control command according to the data rule information of the edge layer, and sends the edge layer control command to the execution class device, To implement edge layer control of the execution class device.
The system of claim 16 wherein at least one of the edge layer fog node, the intermediate layer fog node, and the high level fog node comprises two or more layers of fog nodes.
The system of claim 17, wherein at least one of the edge layer fog node, the middle layer fog node, and the high level fog node comprises two or more layers of fog nodes.