CN-121978980-A - Heterogeneous equipment self-adaptive cooperative control method and system based on edge AI

CN121978980ACN 121978980 ACN121978980 ACN 121978980ACN-121978980-A

Abstract

The invention provides a heterogeneous device self-adaptive cooperative control method and system based on an edge AI, wherein the method comprises the steps of obtaining and analyzing cooperative scene description of a user at an edge gateway side to generate cooperative control intention, generating cooperative executable degree for quantifying cooperative execution feasibility for each virtual service capability, aggregating the cooperative executable degree of each virtual service capability with a preset threshold requirement to form a cooperative execution constraint set, calculating a permission margin of each virtual service capability, taking the minimum permission margin of all the virtual service capabilities as an integral judgment basis to generate cooperative execution permission, triggering and generating corresponding physical devices to execute cooperative action, and immediately blocking execution of the virtual service capability which is not triggered later if the cooperative execution permission fails. The method and the device effectively solve the problem that inconsistent execution and security risk are difficult to manage and control in the complex edge cooperative scene.

Inventors

Zheng Kuixiong
CHEN SHIBIN
HUANG PINFENG
LI JINGHONG
ZHONG YONGHONG

Assignees

广东粤景润科技有限公司

Dates

Publication Date: 20260505
Application Date: 20260408

Claims (10)

1. An edge AI-based heterogeneous device adaptive cooperative control method is characterized by comprising the following steps: s1, acquiring and analyzing collaborative scene description of a user at an edge gateway side, and decomposing the collaborative scene description into a plurality of action-object fragments; matching corresponding virtual service capabilities for each action-object fragment based on a virtual service capability table maintained locally, and encapsulating all the matched virtual service capability sets with a collaborative target description to generate a collaborative control intention; S2, acquiring a candidate device set corresponding to each virtual service capability based on a virtual service capability set in the cooperative control intention, generating cooperative executable degree for quantifying cooperative execution feasibility for each virtual service capability based on the candidate device set, real-time executable state and protocol driving information of the device and combining the candidate device set and driving discrete count, and aggregating the cooperative executable degree of each virtual service capability with a preset threshold requirement to form a cooperative execution constraint set; S3, comparing the cooperative executable degree of each virtual service capability with a threshold requirement according to the cooperative execution constraint set, introducing a risk penalty term, and calculating the permission margin of each virtual service capability; S4, taking the cooperative execution permission and the virtual service capability set as an execution gating signal, traversing each virtual service capability in the virtual service capability set according to a set sequence or a configuration sequence, triggering and generating corresponding physical equipment to execute cooperative action if and only if the cooperative execution permission is permission and the corresponding virtual service capability is not executed, and immediately blocking the execution of the virtual service capability which is not triggered subsequently if the cooperative execution permission is invalid.
2. The heterogeneous device adaptive cooperative control method based on the edge AI of claim 1, wherein the analyzing the cooperative scene description of the user specifically includes: and acquiring a text-form collaborative scene description, expanding the text-form collaborative scene description through a preconfigured action keyword set and an object keyword set, identifying an action fragment and an object fragment from the text through a rule matching mode, and combining the action fragment and the object fragment to form the action-object fragment.
3. The heterogeneous device adaptive cooperative control method based on the edge AI of claim 1, wherein the locally maintained virtual service capability table matches, for each action-object fragment, a corresponding virtual service capability, specifically: and representing the action-object fragment and the virtual service capability as feature sets, evaluating the matching degree by calculating the coincidence degree of the two feature sets, and incorporating the virtual service capability of which the matching result meets a preset threshold into the cooperative control intention.
4. The heterogeneous device adaptive cooperative control method based on the edge AI of claim 1, wherein the candidate device set is directly given by a capability table, and the executable device set is obtained by device-by-device screening of the candidate device set, wherein a screening condition is that a device is in an executable state and is allowed to enter a control path.
5. The edge AI-based heterogeneous device adaptive cooperative control method of claim 1, wherein the driving discrete count is a discrete count at a protocol driving level.
6. The edge AI-based heterogeneous device adaptive collaborative control method according to claim 1, wherein each collaborative execution constraint of the collaborative execution constraint set includes at least a comparison requirement of virtual service capability identification and collaborative executability.
7. The heterogeneous device adaptive cooperative control method based on the edge AI of claim 1, wherein the real-time executable state is reported by a device state acquisition module, and the protocol driving information is read from a device registry.
8. The edge AI-based heterogeneous device adaptive cooperative control method of claim 1, wherein the principle of generating cooperative execution permissions is as follows: When the capacity coverage is higher, the threshold requirement is lower and the drives are more concentrated, the margin is larger, and when the coverage is lower or the drives are more dispersed, the margin is reduced and the blocking is more likely to be triggered.
9. The heterogeneous device adaptive cooperative control method based on the edge AI of claim 1, wherein the triggering of the virtual service capability is generated by calling through a preset virtual service interface and mapping the call into a control instruction of a corresponding physical device by a protocol driving module.
10. An edge AI-based heterogeneous device adaptive cooperative control system, the system comprising: The intention generating module is used for acquiring and analyzing the collaborative scene description of the user at the edge gateway side and decomposing the collaborative scene description into a plurality of action-object fragments; matching corresponding virtual service capabilities for each action-object fragment based on a virtual service capability table maintained locally, and encapsulating all the matched virtual service capability sets with a collaborative target description to generate a collaborative control intention; The constraint generation module is used for acquiring a candidate device set corresponding to each virtual service capability based on the virtual service capability set in the cooperative control intention, generating cooperative executable degree for quantifying cooperative execution feasibility for each virtual service capability based on the candidate device set, real-time executable state and protocol driving information of the device and combining the candidate device set and driving discrete count, and aggregating the cooperative executable degree of each virtual service capability with a preset threshold requirement to form a cooperative execution constraint set; The permission judging module is used for comparing the cooperative executable degree of each virtual service capability with a threshold requirement according to the cooperative execution constraint set, introducing a risk penalty term, and calculating the permission margin of each virtual service capability; And the execution control module is used for traversing each virtual service capability in the virtual service capability set according to a set sequence or a configuration sequence by taking the cooperative execution permission and the virtual service capability set as an execution gating signal, triggering and generating corresponding physical equipment to execute cooperative action if and only if the cooperative execution permission is permission and the corresponding virtual service capability is not executed, and immediately blocking the execution of the virtual service capability which is not triggered subsequently if the cooperative execution permission fails.

Description

Heterogeneous equipment self-adaptive cooperative control method and system based on edge AI Technical Field The invention belongs to the field of intelligent control of building energy, and particularly relates to a heterogeneous equipment self-adaptive cooperative control method and system based on an edge AI. Background With the popularization of applications such as the internet of things, intelligent manufacturing, intelligent buildings and the like, more and more physical devices enter the same network in a local access mode and participate in automatic control, but the devices have remarkable differences in communication protocols, control interfaces, capability granularity, response time sequences and credibility, so that the contradiction of 'getting on, poor use and unstable control' is caused for a long time in engineering landing by cooperative control of multiple devices. In the existing scheme, a rule engine, a flow arrangement or a centralized scheduling strategy is adopted, a plurality of equipment actions are triggered to achieve a linkage effect according to preset conditions, and part of the system further introduces cloud coordination or cloud side intelligent decision to improve scene coverage. However, when a scene is developed from single-point linkage to a cross-device, cross-protocol, cross-room or cross-region collaborative action chain, the online state, controllable state and occupied state of the device change at any time, and the same type of capability can be driven by multiple protocols, collaborative scheduling faces the problems of inconsistent interfaces, unpredictable time sequence, local failure propagation and the like, in the physical world, control actions often have non-rollback performance, and once the control actions are triggered and executed under the condition that conditions are not met or the states are inconsistent, safety risks, energy waste or system instability are easily caused. Meanwhile, the edge side deployment requires low time delay and local autonomy, and is difficult to rely on frequent cloud interaction to complete decision closed loop, the existing protection is concentrated on the access authentication and authority configuration level, and the lack of a unified constraint mechanism for bringing real-time executability and heterogeneous complexity of equipment into a collaborative execution inlet leads to inconsistent results of 'partial equipment is already operated and other equipment is not operated due to unavailability' in collaborative execution, or continues to promote subsequent operation after the state of the equipment is changed, so that the influence range is enlarged. Therefore, how to build an expandable cooperative control mechanism for heterogeneous devices on the edge side enables a cooperative scene to run locally with low time delay, forms a consistency judgment and executable boundary before execution, and can timely block subsequent actions according to real-time state changes in the execution process, so that inconsistent or unsafe cooperative behavior is prevented from falling to the ground, and the key problem to be solved in the current technical system is urgent. Disclosure of Invention The invention aims to provide a heterogeneous equipment self-adaptive cooperative control method and system based on an edge AI, which solve the problems. In order to achieve the above object, in a first aspect of the present invention, there is provided an edge AI-based heterogeneous device adaptive cooperative control method, the method comprising the steps of: s1, acquiring and analyzing collaborative scene description of a user at an edge gateway side, and decomposing the collaborative scene description into a plurality of action-object fragments; matching corresponding virtual service capabilities for each action-object fragment based on a virtual service capability table maintained locally, and encapsulating all the matched virtual service capability sets with a collaborative target description to generate a collaborative control intention; S2, acquiring a candidate device set corresponding to each virtual service capability based on a virtual service capability set in the cooperative control intention, generating cooperative executable degree for quantifying cooperative execution feasibility for each virtual service capability based on the candidate device set, real-time executable state and protocol driving information of the device and combining the candidate device set and driving discrete count, and aggregating the cooperative executable degree of each virtual service capability with a preset threshold requirement to form a cooperative execution constraint set; S3, comparing the cooperative executable degree of each virtual service capability with a threshold requirement according to the cooperative execution constraint set, introducing a risk penalty term, and calculating the permission margin of each virtual service capabil