CN-122021695-A - Rail transit equipment operation and maintenance method and device based on multiple agents

CN122021695ACN 122021695 ACN122021695 ACN 122021695ACN-122021695-A

Abstract

The embodiment of the application discloses a rail transit equipment operation and maintenance method and device based on multiple intelligent agents. The technical scheme provided by the embodiment of the application is that the user intention is generated by responding to the operation and maintenance information input by a user, carrying out intention recognition based on the operation and maintenance information, an agent analysis group is built based on the user intention, the user intention is input into the agent analysis group, the analysis result corresponding to the user intention is output, a task arrangement scheme is generated based on the analysis result, the operation and maintenance task of the task arrangement scheme is executed by at least one pre-built executor agent is called, the execution result of the executor agent is summarized and output, and the executor agent is used for executing the operation and maintenance task of a pre-defined type. By adopting the technical means, the multidimensional analysis information can be fused, the automation processing from the intention recognition to the execution of the operation and maintenance task is realized, and the timeliness of operation and maintenance response, the accuracy of decision and the operation and maintenance safety are improved.

Inventors

Ai Zhenpeng
Luo Tingyu
XIAO JINGHUA
XIE KANG
WANG YUEMIAO
LIU JIA

Assignees

广州佳都智通科技有限公司
佳都科技集团股份有限公司
广东华之源信息工程有限公司
广州华佳软件有限公司
广州佳都城轨智慧运维服务有限公司

Dates

Publication Date: 20260512
Application Date: 20251225

Claims (10)

1. The rail transit equipment operation and maintenance method based on the multiple agents is characterized by comprising the following steps of: responding to operation and maintenance information input by a user, carrying out intention recognition based on the operation and maintenance information, and generating user intention; An agent analysis group is built based on the user intention, the user intention is input into the agent analysis group, an analysis result corresponding to the user intention is output, the agent analysis group comprises analysis agents with a plurality of predefined function roles, and the analysis agents are used for carrying out operation and maintenance analysis on the user intention; And generating a task arrangement scheme based on the analysis result, calling at least one pre-constructed executor agent to execute the operation and maintenance task of the task arrangement scheme, summarizing the execution result of the executor agent and outputting the execution result, wherein the executor agent is used for executing the operation and maintenance task of a predefined type.
2. The multi-agent based rail transit facility operation and maintenance method of claim 1, wherein the performing intent recognition based on the operation and maintenance information, generating user intent, comprises: and invoking a specified base big model to perform information understanding and expansion on the operation and maintenance information to generate the structured user intention, wherein the base big model comprises at least one of a language big model, a video model, a time sequence model and an image model.
3. The multi-agent based rail transit facility operation and maintenance method of claim 1, wherein the inputting the user intention into the agent analysis group and outputting an analysis result corresponding to the user intention comprises: inputting the user intention and the set analysis prompt information into the intelligent body analysis group, sequentially carrying out multi-round operation and maintenance analysis information interaction based on each intelligent body analysis of the intelligent body analysis group, and outputting an analysis result corresponding to the user intention.
4. The multi-agent based rail transit facility operation and maintenance method of claim 3, wherein the analysis agents include operation and maintenance technology analysis agents, historical data analysis agents, and security management agents; The operation and maintenance technology analysis agent is used for accessing a pre-constructed operation and maintenance knowledge base, generating operation and maintenance technology suggestions based on operation and maintenance knowledge information provided by the operation and maintenance knowledge base, and outputting the operation and maintenance technology suggestions as operation and maintenance analysis information; The historical data analysis agent is used for inquiring a historical data interface of the pre-constructed operation and maintenance tool library, acquiring corresponding historical operation and maintenance data, performing operation and maintenance analysis based on the historical operation and maintenance data, and outputting corresponding operation and maintenance analysis information; the security management and control agent is used for accessing a pre-constructed operation and maintenance knowledge base, performing security risk assessment based on operation and maintenance knowledge information provided by the operation and maintenance knowledge base, generating a security risk assessment result, and outputting the security risk assessment result as operation and maintenance analysis information.
5. The multi-agent based rail transit facility operation and maintenance method of claim 1, wherein the generating a mission orchestration scheme based on the analysis results comprises: Inquiring a pre-configured workflow template library based on the analysis result, and acquiring a workflow template matched with the analysis result from the workflow template library, wherein the workflow template comprises executor agent types, execution sequences and input information of a plurality of task flows; And creating a task arrangement scheme corresponding to the operation and maintenance information according to the workflow template.
6. The multi-agent based rail transit facility operation and maintenance method of claim 1, wherein the executor agents are configured to invoke a tool interface of a pre-built operation and maintenance tool library to perform predefined types of operation and maintenance tasks including at least one of facility inspection, fault diagnosis, and personnel scheduling.
7. Rail transit equipment operation and maintenance device based on multiple agents, characterized by comprising: The intention recognition module is used for responding to the operation and maintenance information input by the user, carrying out intention recognition based on the operation and maintenance information and generating the intention of the user; The operation and maintenance analysis module is used for constructing an agent analysis group based on the user intention, inputting the user intention into the agent analysis group, and outputting an analysis result corresponding to the user intention, wherein the agent analysis group comprises analysis agents with a plurality of predefined function roles, and the analysis agents are used for carrying out operation and maintenance analysis on the user intention; and the task execution module is used for generating a task arrangement scheme based on the analysis result, calling at least one pre-constructed executor agent to execute the operation and maintenance task of the task arrangement scheme, summarizing the execution result of the executor agent and outputting the execution result, wherein the executor agent is used for executing the operation and maintenance task of a predefined type.
8. Rail transit equipment operation and maintenance system based on multiple agents, characterized by comprising: the base large model is used for understanding and generating multi-mode data; the operation and maintenance knowledge base is used for storing operation and maintenance knowledge information associated with the operation and maintenance of the rail transit equipment; the operation tool library is used for providing an operation tool interface for interacting with the operation platform; And the plurality of agents are used for calling the base large model, the operation and maintenance knowledge base and the operation and maintenance tool base to cooperatively execute the rail transit equipment operation and maintenance method based on the plurality of agents according to any one of claims 1-6.
9. An electronic device, comprising: a memory and one or more processors; the memory is used for storing one or more programs; The one or more programs, when executed by the one or more processors, cause the one or more processors to implement the multi-agent based rail transit facility operation and maintenance method of any one of claims 1-6.
10. A storage medium containing computer executable instructions which, when executed by a computer processor, are for performing the multi-agent based rail transit facility operation and maintenance method of any one of claims 1-6.

Description

Rail transit equipment operation and maintenance method and device based on multiple agents Technical Field The embodiment of the application relates to the technical field of computers, in particular to a rail transit equipment operation and maintenance method and device based on multiple intelligent agents. Background At present, in the operation and maintenance scene of the rail transit equipment, a cooperative working mode based on an operation and maintenance platform is generally adopted, a plurality of roles such as an maintainer, a worker, a technical engineer, a safety manager, a project manager and the like participate, and the processes such as fault report, work order dispatch, maintenance execution and the like are completed based on a platform function module (such as inspection management, maintenance management, scheduling management and the like). Although the method improves the working efficiency compared with the traditional paper work order and telephone notification and other working methods, the method still has the problems of lag response, long flow, outstanding inter-system data barriers, decision dependence on manual experience, insufficient safety risk identification and the like. Therefore, the existing technical scheme for assisting equipment operation and maintenance by adopting the intelligent agent can realize automation of operation and maintenance flow in the aspects of equipment state monitoring, fault diagnosis and the like, thereby simplifying the operation and maintenance flow and avoiding the problems of decision errors, response lag and the like caused by relying on manual experience. However, when the operation and maintenance task is executed by the existing scheme adopting the operation and maintenance of the intelligent agent auxiliary equipment, a single intelligent agent is difficult to comprehensively adapt to diversified work demands, and when a cross-system decision is involved, the output result of the intelligent agent auxiliary equipment has the conditions of one-sidedness or missing key factors and the like, so that the overall work efficiency and effect are affected. In addition, when a single agent processes tasks in different fields, context information in different fields needs to be frequently switched, so that output of the single agent is unstable and even wrong, and global coordination and efficient optimization are difficult to achieve. Disclosure of Invention The embodiment of the application provides a rail transit equipment operation and maintenance method and device based on multiple agents, which can be used for fusing multidimensional analysis information, realizing the automatic processing from intention recognition to execution of operation and maintenance tasks, and solving the technical problem that an output result is unstable due to frequent analysis of one-sided and context switching of a single agent in cross-system decision. In a first aspect, an embodiment of the present application provides a rail traffic equipment operation and maintenance method based on multiple agents, including: responding to operation and maintenance information input by a user, carrying out intention recognition based on the operation and maintenance information, and generating user intention; An agent analysis group is built based on the user intention, the user intention is input into the agent analysis group, an analysis result corresponding to the user intention is output, the agent analysis group comprises analysis agents with a plurality of predefined function roles, and the analysis agents are used for carrying out operation and maintenance analysis on the user intention; And generating a task arrangement scheme based on the analysis result, calling at least one pre-constructed executor agent to execute the operation and maintenance task of the task arrangement scheme, summarizing the execution result of the executor agent and outputting the execution result, wherein the executor agent is used for executing the operation and maintenance task of a predefined type. Further, the performing intention recognition based on the operation and maintenance information, generating a user intention, includes: and invoking a specified base big model to perform information understanding and expansion on the operation and maintenance information to generate the structured user intention, wherein the base big model comprises at least one of a language big model, a video model, a time sequence model and an image model. Further, the inputting the user intention into the agent analysis group, outputting an analysis result corresponding to the user intention, includes: inputting the user intention and the set analysis prompt information into the intelligent body analysis group, sequentially carrying out multi-round operation and maintenance analysis information interaction based on each intelligent body analysis of the intelligent body analysis group, and outputting a