Research On Cooperative Adaptive Learning Control Methods For Multi-agent System

In recent years,with the wide application of cooperative control of multi-agent systems in many fields,more and more researchers have paid attention to it.At present,in the field of distributed control of multi-agent systems,the combination of adaptive learning technology and optimal control method has made great progress.At present,the optimal cooperative control method of multi-agent systems based on adaptive learning mainly solves the cooperative control problem of multi-agent systems without system faults,and only discusses the consensus control method under fixed communication topology.Based on the above discussion,this paper studies the cooperative fault-tolerant control design of multi-agent systems based on adaptive learning method.Firstly,the problem of cooperative optimal fault-tolerant control for data-driven linear multi-agent systems with fixed directed communication topology is studied.By defining a new cooperative quadratic performance index,a distributed optimal fault-tolerant control algorithm is designed,and the necessary and sufficient conditions for optimal control to minimize the quadratic performance index are established by using algebraic Riccati equation.especially when the dynamic matrix of the multi-agent subsystem is unknown,an iterative algorithm is designed to calculate the feedback gain of the system by using adaptive dynamic programming technology.Furthermore,a cooperative optimal fault-tolerant control algorithm is designed to ensure that the multi-agent system achieves the control goal of consistent tracking.Finally,a simulation example is given to verify the effectiveness of the proposed control method.Furthermore,the optimal fault-tolerant control problem of linear multi-agent systems under network intermittent denial of service attacks is studied.Assuming that the communication network of multi-agent is a time-varying asymmetric topology,a new resilient optimal cooperative fault-tolerant control strategy is proposed by using the local information of the agent unit,and the cooperative quadratic performance index of the multi-agent system is optimized.A self-learning iterative algorithm using local system state and input information is proposed to solve the algebraic Riccati equation,and then the feedback controller gain of the subsystem is calculated to achieve the goal of resilient cooperative fault-tolerant tracking control.