CN-120560041-B - Safety boundary consistency control method for multi-agent space-time dynamic system

Abstract

The invention discloses a safety boundary consistency control method of a multi-agent space-time dynamic system, which comprises the following technical steps of firstly constructing a leader-follower space-time dynamic mathematical model of the multi-agent system based on a partial differential equation, then constructing an abnormal working condition mathematical model comprising an actuator fault and a network spoofing attack, deducing and obtaining an error dynamic system equation by defining a system consistency control target, constructing an energy function by using a Lyapunov stability theory, deducing and obtaining a system stability criterion, and finally designing a safety boundary consistency control method to realize consistency control of the space-time dynamics of the multi-agent system. The method effectively solves the problem of high cost caused by the arrangement of the executors in the whole space domain in the traditional scheme, and the problem that the multi-agent system is easy to be interfered by network attack in the communication process, and has remarkable application value in engineering practice.

Inventors

• WANG ZIPENG
• WANG BOWEN
• QIAO JUNFEI
• HAN HONGGUI
• LI FANGYU

Assignees

• 北京工业大学

Dates

Publication Date

20260508

Application Date

20250711

Claims (3)

1. 1. A method for controlling the consistency of safety boundaries of a multi-agent space-time dynamic system, the method comprising; Step one, constructing a multi-agent space-time dynamic system model based on a parabolic partial differential equation, wherein the multi-agent space-time dynamic system comprises Agents, each agent numbered The leader labels the agent as l, and the leader agent model in the multi-agent space-time dynamic system is as follows: ; Wherein, the Indicating that the leader agent is at time t State of a site, spatial variable Time variable The initial state of the leader agent is Is a positive definite matrix, a is a known constant real matrix; the mathematical model of a single follower agent in the multi-agent space-time dynamic system is as follows: ; Wherein, the Indicating that the ith agent is at time t In the state of the device, the device is in a state, The initial values of N follower agents are expressed as A control input indicative of an occurrence of an actuator failure; step two, establishing a space-time representation of the faults of the executors of part of the intelligent agents at the system boundary and a mathematical model of the intelligent agents suffering from deception attack in data transmission; Step three, defining a leader-follower consistency error to obtain a consistency state error system; designing a spoofing attack and execution failure safety boundary consistency controller; constructing a corresponding Lyapunov function to obtain a condition for realizing safety boundary consistency control of the system; during operation, part of the intelligent agents can generate actuator faults, and the mathematical description is as follows: ; Wherein, the Representing an actuator influence factor of an ith agent, the influence factor varying with time t, assuming the influence factor Is bounded and satisfies the following inequality , wherein, And Respectively a lower bound and an upper bound of the influencing factor, in addition to A control input representing an ith agent; comprehensively analyzing the influence of the spoofing attack and the actuator fault, designing a boundary control protocol as follows: ; Where k is the control gain, Is a communication weight matrix between the intelligent agents, if communication exists between the intelligent agent j and the intelligent agent i, the communication weight matrix is formed Otherwise For the coupling weight coefficient between the agent i and the leader, when When the intelligent agent i and the leader can conduct data transmission When the data transmission between the intelligent agent i and the leader is impossible, , For the laplace matrix of the system, , Is an input degree matrix of the system, 。
2. 2. The method for controlling the consistency of the safety boundaries of a multi-agent space-time dynamic system according to claim 1, wherein the agents in the multi-agent space-time dynamic system need to perform frequent interactions through a network and are vulnerable to network attacks, and a spoofing attack is considered, wherein the spoofing attack is a form of network attack for reducing the reliability of information transfer by tampering with data transmitted over a wireless communication network between agents, and wherein an attacker performs tampering with communication data by intercepting data signals from agents to agents ready to be transmitted through the network and transforming the intercepted data by using a nonlinear function, and the mathematical expression is as follows: ; Wherein, the Is a nonlinear function satisfying the following inequality: F is a nonlinear function And (3) a constant matrix determined.
3. 3. The multi-agent spatiotemporal dynamic system safety margin consistency control method of claim 1, wherein a leader-follower consistency error is defined The error system is obtained as follows: ; ; for any initial state, the leader-follower multi-agent spatiotemporal dynamic system is said to achieve the desired consistency control if the error signal satisfies the following equation relationship: 。

Description

Safety boundary consistency control method for multi-agent space-time dynamic system Technical Field The invention relates to the field of intelligent control, in particular to a safety boundary consistency control method of a multi-agent space-time dynamic system. Background The multi-agent system is a distributed network system composed of a plurality of mutually cooperative or competing agents, which have certain intelligence and integrate technologies such as control, optimization and the like. Through the information interaction between the intelligent agents, the system can complete complex tasks in the fields of aerospace, national defense, industry and the like. In recent years, multi-agent systems have been widely used in different scenarios. Thus, the problem of cooperative control of multi-agent systems has received extensive attention from numerous researchers. Cooperative control of the multi-agent system is capable of accomplishing a number of different types of tasks including consistency, formation control, cluster control, beehive control, network optimization, and the like. The consistency problem is the most basic and key problem in the coordination control of the multi-agent system, and has important significance for ensuring the stability and reliability of the system. Multi-agent systems have wide application in the industrial and military fields where stability and safety of control are critical, and any mistake can lead to significant economic losses. For example, multi-agent systems are widely used for collaborative assembly of robotic arms in automotive manufacturing lines. If communication or cooperation between the mechanical arms is problematic, stagnation of the whole production line may be caused, and thus, huge economic loss is caused. Therefore, research on the safety consistency control of the multi-agent system has important significance for ensuring the stability and safety of the system-! Current research in the field of multi-agent system safety control presents a significant modeling paradigm shift trend. The traditional research mainly adopts a model building method of a normal differential equation (ODE), and the model describes system evolution through a single-dimensional function of time variable, but can effectively describe time domain dynamic coupling among intelligent agents, but cannot characterize the spatial distribution characteristic and space-time coupling effect of the system. The modeling mode essentially simplifies the multi-intelligent system with the space expansion characteristic into a particle model, so that the inherent space-time cooperative characteristic of the system is filtered in the modeling process, and the accurate modeling and analysis of the intelligent behaviors of the group in a complex scene are severely restricted. With respect to this theoretical limitation, modeling methods based on Partial Differential Equations (PDEs) have achieved breakthrough progress in recent years. The PDE model can simultaneously represent continuous evolution rules of the intelligent agent in time and space double dimensions by introducing space dimension variables, and is particularly suitable for describing typical multi-intelligent agent systems such as unmanned aerial vehicle clusters with space distribution characteristics, distributed sensor networks and the like. At present, scholars have constructed various space-time coupling models such as reaction-diffusion type PDE, hyperbolic type PDE and the like, research on key problems such as cooperative control, formation optimization and the like is carried out on the basis, and achievement with milestone significance is achieved in aspects such as space-time consistency control, distributed state estimation and the like. It is worth noting that the existing multi-agent space-time dynamic system safety consistency research mostly adopts an intra-domain control strategy, namely, a distributed controller or an actuator is arranged inside a system space domain. In contrast, the invention provides a boundary consistency control method, which realizes the cooperative control of a global system by designing the control input of an intelligent agent on a space boundary. The control mode only needs to exert control action at the limited boundary of the system space domain, so that the deployment density of the executor and the hardware cost are greatly reduced, and the boundary controller is easier to physically deploy and has stronger anti-interference capability in the engineering realization level. In recent years, some researchers have primarily studied the problem of consistency control of multi-agent space-time dynamic systems and have achieved some meaningful research results. But the problem of safety margin consistency control for multi-agent spatiotemporal dynamic systems under spoofing attacks and actuator failures has not been considered so far. Disclosure of Invention The invention provides a mu