CN-121985015-A - Industrial Internet of things agent virtual-real coordination dynamic hierarchical decision response method, system, terminal equipment and medium

CN121985015ACN 121985015 ACN121985015 ACN 121985015ACN-121985015-A

Abstract

The invention discloses an intelligent virtual-real coordination dynamic hierarchical decision response method, an intelligent virtual-real coordination dynamic hierarchical decision response system, terminal equipment and medium for an industrial Internet of things, and relates to the technical field of intelligent control of the industrial Internet of things. The method comprises the steps of obtaining real-time data flow of industrial Internet of things terminal equipment and carrying out structural processing to obtain structural task data, classifying tasks into simple tasks or complex tasks through a stage perceptron, generating candidate actions and pre-execution candidate sets through an edge lightweight inference model by the simple tasks and determining final control instructions, generating main actions through a core inference model by the complex tasks, matching or directly generating the final control instructions through the pre-execution candidate sets, and controlling the terminal equipment to execute decision response through the final control instructions. The invention realizes the accurate processing of the task branch paths, multiplexes the pre-execution control instructions, greatly shortens the response delay, improves the resource utilization efficiency, and ensures the safety and reliability of the industrial scene decision.

Inventors

LU WEICHAO

Assignees

深圳开鸿数字产业发展有限公司

Dates

Publication Date: 20260505
Application Date: 20251226

Claims (17)

1. The utility model provides an industry thing networking agent virtual reality is dynamic hierarchical decision response method in coordination, its characterized in that, the method includes: Acquiring a real-time data stream of industrial Internet of things terminal equipment, and carrying out structuring treatment on the real-time data stream to obtain structured task data; classifying tasks by a phase perceptron based on the structured task data, wherein the tasks are classified as simple tasks or complex tasks; Generating a main action through a core reasoning model for the complex task, and obtaining a final control instruction through matching or direct generation of the pre-execution candidate set; and controlling the industrial Internet of things terminal equipment to execute decision response operation aiming at the task through the final control instruction.
2. The method for responding to real-time hierarchical decision of industrial internet of things agent virtual-real coordination according to claim 1, wherein the step of obtaining the real-time data stream of the industrial internet of things terminal device comprises the following steps: Collecting real-time operation data through a sensor and a controller of the industrial Internet of things terminal equipment; And carrying out noise reduction pretreatment on the acquired real-time operation data to obtain the real-time data stream.
3. The method for responding to real-time hierarchical decision of industrial internet of things agent virtual-real coordination according to claim 1, wherein the structuring the real-time data stream to obtain structured task data comprises: Analyzing the real-time data stream through an industrial protocol, and extracting key feature vectors; And executing feature standardization processing on the key feature vector to obtain structured task data, wherein the structured task data is used for expressing task content to be processed.
4. The method for responding to the dynamic hierarchical decision of the virtual-real coordination of the industrial internet of things agent according to claim 1, wherein the step of classifying the task by the stage perceptron based on the structured task data comprises the steps of: Calculating task complexity based on the structured task data; Comparing the task complexity to a dynamic threshold; if the task complexity is below the dynamic threshold, the task is classified as a simple task, otherwise the task is classified as a complex task.
5. The method for responding to the dynamic hierarchical decision of the virtual-real coordination of the industrial internet of things agent according to claim 4, wherein the structured task data at least comprises a decision time limit, a device risk level and a data dimension, the task complexity is calculated based on the structured task data, and the method comprises the following steps: and dynamically calculating the task complexity based on decision time limit, equipment risk level and data dimension in the structured task data.
6. The method for responding to the dynamic hierarchical decision of the virtual-real coordination of the industrial internet of things agent according to claim 4, wherein the dynamic threshold is calculated by a preset basic threshold and a device risk level in the structured task data, and the device risk level is positively related to the dynamic threshold.
7. The method for responding to the dynamic hierarchical decision of the virtual-real coordination of the industrial internet of things agent according to claim 1, wherein after the task is classified by the stage perceptron based on the structured task data, the method further comprises: And screening to obtain a release task through a preset screening rule based on the current load and the expected benefits of the task, wherein the release task is a task allowing action generation, and the expected benefits of the task are comprehensively judged by action overlapping, time saving, hit probability, transmission overhead and calculation overhead.
8. The method of claim 1, wherein the edge lightweight inference model is an inference model deployed at an edge node or a local of the industrial internet of things, and is obtained by model structure simplification and parameter clipping optimization, and the generating a plurality of candidate actions and pre-execution candidate sets by the edge lightweight inference model for the simple task, and determining a final control instruction comprise: Generating a preset number of candidate actions by using an edge lightweight inference model based on structured task data corresponding to the simple task, wherein the candidate actions are used for responding to the requirements of the task; Executing pre-execution operation on all candidate actions in a preset simulation environment to generate a corresponding number of control instructions and simulation running results; Caching the candidate actions, the corresponding control instructions and the simulation running results to obtain a pre-execution candidate set in the cache; And determining a final control instruction in the candidate actions and the corresponding control instructions based on a preset scoring mechanism.
9. The method for responding to the virtual-real cooperative dynamic hierarchical decision of the industrial internet of things agent according to claim 8, wherein the caching the candidate actions, the corresponding control instructions and the simulation running results to obtain the pre-execution candidate set in the cache comprises: And performing associated storage on the candidate actions, the corresponding control instructions and the simulation operation results, and associating life cycle aging of the industrial Internet of things terminal equipment to form a pre-execution candidate set, wherein the pre-execution candidate set is stored by adopting a time sequence database label.
10. The method for responding to the virtual-real coordination dynamic hierarchical decision of the industrial internet of things agent according to claim 8, wherein determining the final control instruction in the candidate actions and the corresponding control instructions based on a preset scoring mechanism comprises: Obtaining a similarity score through semantic similarity evaluation based on the candidate actions and the targets of the tasks; and selecting the control instruction corresponding to the candidate action with the highest similarity score as the final control instruction.
11. The method of claim 1, wherein the core inference model is a high-precision inference model with a complete inference architecture for processing complex tasks, the high-performance inference model is deployed in a cloud or an industrial internet of things edge node, the main action is generated by the core inference model for the complex tasks, and the final control instruction is obtained by the pre-execution candidate set matching or direct generation, and the method comprises the following steps: based on the structured task data corresponding to the complex task, generating a main action by reasoning through the core reasoning model; verifying whether the master action exists in the pre-execution candidate set; If the main action exists, taking a control instruction corresponding to the main action in the pre-execution candidate set as the final control instruction; If not, generating the final control instruction based on the master action.
12. The method for responding to the virtual-real coordination dynamic hierarchical decision of the industrial internet of things agent according to claim 11, wherein before the step of generating the initiative by reasoning through the core reasoning model based on the structured task data corresponding to the complex task, the method further comprises: And dynamically assigning tasks to the edge nodes through a distributed soft bus based on the computational power demand vector of the complex task, and preferentially distributing computing resources of the edge nodes to provide computational power, wherein the distributed soft bus realizes low-delay coordination among the edge devices through a network topology distance optimization mechanism.
13. The method for responding to the industrial internet of things agent virtual-real cooperative dynamic hierarchical decision according to claim 11, wherein the taking the control instruction corresponding to the main action in the pre-execution candidate set as the final control instruction includes: Based on a preset industrial protocol white list and a hardware signature, executing compliance verification on a control instruction corresponding to the main action in the pre-execution candidate set and a simulation running result; And if the verification is passed, taking the control instruction corresponding to the main action as a final control instruction.
14. The method for responding to the dynamic hierarchical decision of the virtual-real coordination of the industrial internet of things agent according to claim 10, further comprising, after obtaining the final control instruction: aiming at a simple task, obtaining a similarity score of a candidate action corresponding to a final control instruction and a target of the task; Comparing the similarity score with a switching threshold, wherein the switching threshold is calculated by a preset basic threshold and the structured task data; And if the similarity score is lower than the switching threshold, switching the simple task into a complex task.
15. An industrial internet of things agent virtual-real coordination dynamic hierarchical decision response system, which is characterized by comprising: the system comprises a structured task data acquisition module, a data processing module and a data processing module, wherein the structured task data acquisition module is used for acquiring a real-time data stream of industrial Internet of things terminal equipment and carrying out structured processing on the real-time data stream to obtain structured task data; The task classification module is used for classifying tasks through a stage sensor based on the structured task data, wherein the tasks are classified into simple tasks or complex tasks; The final control instruction generation module is used for generating a plurality of candidate actions and pre-execution candidate sets through an edge lightweight inference model according to the simple task and determining a final control instruction; And the decision response module is used for controlling the industrial Internet of things terminal equipment to execute decision response operation aiming at the task through the final control instruction.
16. A terminal device, characterized in that the terminal device comprises a memory, a processor and an industrial internet of things agent virtual-real cooperative dynamic hierarchical decision response program stored in the memory and capable of running on the processor, wherein the processor realizes the steps of the industrial internet of things agent virtual-real cooperative dynamic hierarchical decision response method according to any one of claims 1-14 when executing the industrial internet of things agent virtual-real cooperative dynamic hierarchical decision response program.
17. A computer readable storage medium, wherein the computer readable storage medium stores an industrial internet of things agent virtual-real coordination dynamic hierarchical decision response program, and when the industrial internet of things agent virtual-real coordination dynamic hierarchical decision response program is executed by a processor, the steps of the industrial internet of things agent virtual-real coordination dynamic hierarchical decision response method according to any one of claims 1-14 are implemented.

Description

Industrial Internet of things agent virtual-real coordination dynamic hierarchical decision response method, system, terminal equipment and medium Technical Field The invention relates to the technical field of intelligent control of industrial Internet of things, in particular to a virtual-real cooperation dynamic hierarchical decision response method, a system, terminal equipment and a medium for an intelligent agent of the industrial Internet of things. Background The industrial Internet of things is widely applied to the fields of intelligent factory production line regulation, energy network scheduling, high-risk operation safety and the like, and the core requirement is to realize quick response and accurate decision of equipment tasks. In the prior art, a serial architecture of sensing, reasoning and executing is mostly adopted, so that decision delay linearly increases along with the number of steps, and the delay of a complex production line regulation task is even more than 30 minutes. Meanwhile, the edge equipment adopts a fixed scheduling strategy, so that the task complexity dynamic change is difficult to adapt, and the system congestion is easy to cause under high load. In addition, the industrial protocol verification is separated from the decision engine, the high-risk scene action execution lacks real-time risk perception, and the safety response lag is caused by the dependence on manual intervention. Therefore, an integrated decision response method capable of reducing task control delay, load balancing and real-time sensing is needed to fill the gap in the prior art. Disclosure of Invention The technical problem to be solved by the invention is that in the field of intelligent control of the industrial Internet of things, the prior art has the disadvantages of low task response efficiency, poor system adaptability and insufficient risk prevention and control in high-risk scenes due to corresponding delay accumulation of decisions, unbalanced resource allocation and lack of a safety mechanism. Therefore, an effective solution is needed to solve the above technical problems. In order to solve the technical problems, the technical scheme adopted by the invention is as follows: in a first aspect, the invention provides an industrial internet of things agent virtual-real coordination dynamic hierarchical decision response method, which comprises the following steps: Acquiring a real-time data stream of industrial Internet of things terminal equipment, and carrying out structuring treatment on the real-time data stream to obtain structured task data; classifying tasks by a phase perceptron based on the structured task data, wherein the tasks are classified as simple tasks or complex tasks; Generating a main action through a core reasoning model for the complex task, and obtaining a final control instruction through matching or direct generation of the pre-execution candidate set; and controlling the industrial Internet of things terminal equipment to execute decision response operation aiming at the task through the final control instruction. In one implementation manner, the acquiring the real-time data stream of the industrial internet of things terminal device includes: Collecting real-time operation data through a sensor and a controller of the industrial Internet of things terminal equipment; And carrying out noise reduction pretreatment on the acquired real-time operation data to obtain the real-time data stream. In one implementation, the structuring the real-time data stream to obtain structured task data includes: Analyzing the real-time data stream through an industrial protocol, and extracting key feature vectors; And executing feature standardization processing on the key feature vector to obtain structured task data, wherein the structured task data is used for expressing task content to be processed. In one implementation, the grading the task by the phase perceptron based on the structured task data includes: Calculating task complexity based on the structured task data; Comparing the task complexity to a dynamic threshold; if the task complexity is below the dynamic threshold, the task is classified as a simple task, otherwise the task is classified as a complex task. In one implementation manner, the structured task data at least includes a decision time limit, a device risk level, and a data dimension, and the calculating, based on the structured task data, task complexity includes: and dynamically calculating the task complexity based on decision time limit, equipment risk level and data dimension in the structured task data. In one implementation, the dynamic threshold is calculated from a preset base threshold and a device risk level in the structured task data, where the device risk level is positively correlated to the dynamic threshold. In one implementation, after the task is classified by the stage perceptron based on the structured task data, the method further includ