Consensus Of Multi-agent Systems Based On Event-triggered Control Strategy

Consensus problem is a crucial research in the cooperative control.Its goal is to achieve an agreement state of all agents by designing appropriate control protocols,which are based on the state information of agents themselves and corresponding neighboring agents.Therefore,the core problem of consensus research is how to design control protocol.Based on event-triggered control,low-gain feedback control,sliding mode control,quantitative control,observation control and adaptive control,this paper studies the consensus and bipartite consensus of multi-agent systems.The first part studies the consensus problem of leaderless multi-agent systems with input saturation.Firstly,basing on the low-gain feedback control method,we design two event-triggered control protocols,namely,centralized event-triggered control protocol and distributed event-triggered control protocol.Secondly,by applying the Lyapunov stability theory,the sufficient criterion for multi-agent systems to reach the consensus is obtained.Finally,it is proved that both event-triggered protocols designed can exclude Zeno behavior,that is,there are no infinite number of triggers in a finite time interval.In the second part,the consensus problem of leader-following multi-agent systems with input saturation is researched.Firstly,based on the output feedback information,the observational system is constructed to estimate the state information of the original system.Secondly,under the designed centralized event-triggered control protocol,the related criteria of leader-following consensus are given by applying Lyapunov stability theory and Cauchy's convergence criterion.Then,it is proved that the event-triggered protocol was designed to avoid the Zeno behavior.Finally,the group centralized event-triggered mechanism is considered.Combining the theoretical results and numerical examples,this protocol has the characteristics of distributed event-triggered protocol.In the third part,the bipartite consensus problem of leader-following multi-agent systems with matched disturbances is studied.Firstly,the observation system is constructed to estimate the state information of the original system,and two novel integral sliding mode surfaces are designed to eliminate the disturbances and the disturbances estimation.Secondly,by designing the adaptive quantization distributed event-triggered control protocol and using the Riccati equation,La Salle's invariance principle and Lyapunov stability theory,we obtain the relevant criteria for achieving bipartite consensus.Finally,it is proved that the adaptive distributed event-triggered protocol designed can exclude Zeno behavior.