Input Time Delay Margin In Event-triggered Consensus Of Multi-agent Systems

Generally,the traditional controlled systems have employed centralized control strategy,namely,the central controllers are uniformly responsible for getting together information of systems and then taking corresponding control actions.When systems become complicated and large-scale,the control method can not be employed effectively in practice: the burden of communication and consumption of resource for agents achieving the consensus will increase,and the control rules will become more and more complicated,which all make the traditional method difficult to implement.Recently,the distributed control has displayed its advantage over traditional central control method in overcoming the faults above,especially in resisting disturbance from external environment.Therefore,it has attracted much attention from researches in various fields.By making full use of traits of distributed control: less consumption of energy resource and local control,researches have applied it to distributed sensor networks for surveillance,unmanned aerial vehicle(UAV)and distributed cooperation optimization and so on.This paper mainly focuses on the event-triggered consensus of general linear multiagent system.Under the conditions that the communication network is directed and the time delay is taken into consideration,we accurately compute the input time delay margin to ensure the consensus of controlled system.Meanwhile,we obtain a more general condition to exclude the Zeno behavior,that is to say,the number of communication between agents determined by the triggering function can not be infinite within a finite time.Next,a self-triggered control scheme is proposed for agents to avoid continuously monitoring themselves state,also,considering that the agent states are not measurable in practice,we design observer-based control scheme to deal with this problem.At last,simulations are presented to demonstrate the theoretical results above.This paper mainly composes of five parts:(1)We introduce the context and significance of event-triggered consensus of general linear multi-agent system,and the existing research findings about it,which are certain to give a direction for our work.By presenting the research above,we derive our new results which to some extent greatly improve the existing results.(2)For the consensus of multi-agent system,the notations and definition are introduced.Next,controller containing time delay is proposed and lemmas which this paper concerns are presented.Procedure of setting parameters is given for our convenience to choose appropriate parameters.(3)Based on the preparations above,allowable time delay margin is calculated.Especially,by decomposing high dimensional system into several low dimensional subsystems,the time delay margin is easy to obtain.Finally,the Zeno behavior is excluded,meanwhile a more general condition to exclude Zeno behavior is derived compared to some existing results.(4)To avoid the continuous monitoring for agent states,and considering that the states of agents are not available in practice,self-triggered control scheme and observerbased control scheme are introduced,respectively.Self-triggered control scheme means that agents can forecast next triggering instants at current triggering instant without continuously monitoring agents state,this saves a large number of energy.Observer-based control scheme only uses the output information to control multi-agent system,meanwhile the time delay margin is also obtained.(5)Simulations are given to verify the theoretical results.Concretely,time delay margin of system,Zeno behavior and the parameters effect on the performance of system are studied.Finally,all the work of this paper is summarized,and the need of the work is prospected.