Research On Event-triggered Formation Control Of Multi-agent Systems

The multi-agent system is composed of several intelligent agents with sensing,computing,execution and communication abilities.Compared with the individual agent,multi-agent cooperation can improve the ability and efficiency of task execu-tion,reduce the cost of individual,increase system redundancy and risk resistance.So more and more researchers pay attention to cooperation control of multi-agent systems.The formation control is one of those important branches of multi-agent cooperative control.The goal of formation control is to make multiple agents keep or change their formation in the process of movement according to the task needs.It has a wide range of applications in military and civil fields,such as battle-plane formation,satellite formation,UAV formation performance and so on.In actual sys-tems,because of the limited communication bandwidth and computing resources,and the requirement of forming the formation within specified time,it is of great practical significance to study and improve the convergence speed of the formation control strategy and save the limited computing and communication resources.This paper mainly studies the event-triggered formation control problem of multi-agent systems,and designs different distributed control strategies for some typical system models and control targets,including robust control,event-triggered control,self-triggered control,finite-time/fixed-time control and so on.The main research content of the thesis includes the following aspects:1.Aiming at the time-varying formation problem of multi-agent systems,the observer-based event-triggered control strategy and self-triggered control strategy are designed respectively.By introducing the state observation error into the event-triggered function,the convergence speed of the state observer and the formation accuracy of the system can be improved.The two control strategies can achieve time-varying formation control and avoid Zeno behavior,while the self-triggered strategy can avoid continuous communication between agents.2.For the disturbed multi-agent system,the formation tracking control prob-lem with a leader is studied.Fixed-time event-triggered and self-triggered control strategies are designed respectively.Fixed-time control can effectively improve the convergence rate of the system,and the convergence time is independent of the ini-tial state of the system.Event-triggered control can reduce the number of controller updates,and self-triggered control can avoid continuous communication between agents.3.For the sampled-data-based multi-agent system,the formation tracking con-trol of the first-order multi-agent system based on sampling data is studied.The limited time event-triggered and self-triggered strategy are designed respectively,which can avoid continuous communication.4.For the sampled-data-based nonlinear high-order multi-agent system,the formation control problem is studied,and the distributed dynamic event-triggered mechanism is designed.By dynamically adjusting the trigger function parameters,the system trigger times can be effectively reduced.5.At present,the research of multi-agent cooperative control mainly focuses on theoretical research and simulation verification.In this paper,the multi-agent formation control strategy is verified by using the multi unmanned vehicle experi-mental platform,and the finite-time event-triggered formation tracking experiment and fixed-time event-triggered formation tracking experiment based on sampling are designed respectively.The effectiveness of finite-time control,fixed-time con-trol,event-triggered control and self-triggered control are verified by experiments.