Consensus Of Multi-Agent System Via Event-Triggered Impulsive Control

In recent years,the research of multi-agent system has attracted wide attention,which is widely used in the intersection of disciplines related to control,and has become the research hotspot in the field of control and computer.Among the many research problems of multi-agent system,the consensus problem is the most predominant one.With the development of consistency research,the traditional consensus problem has been unable to cope with the changes and needs of the actual environment.Group consensus has become a new research hotspot.In order to solve the consensus problem of multi-agent system more efficiently,there are many novel control methods,the event-triggered impulsive control method is a more effective modern control method which is improved on the event triggering method,and has become a research hotspot in the control method.Aim at the group consensus problem of general linear time-invariant system with or without leder agent,and the complete consensus condition of multi-agent system under time-varying system is also studied.The research content is to design and analyze the event-triggered impulsive controller for the system.Which will be introduced from the following four parts:The background and significance of the research topic is introduced in the first part,and state the importance of the research topic.Including the related theoretical knowledge and development overview of the research topic,and the basic concepts such as graph theory and symbolic representation.This part plays an important and fundamental role in the consensus research of multi-agent system.The group consensus problem of linear multi-agent systems is studied in the second part,so as a novelty event-triggered controller and impulsive controller combined control protocol will be designed,and any nonzero projection parameters will be introduced to the system model.It means that the results is a generalization of the completely consensus and the anti-binary group consensus form.By constructing the Lyapunov function and using algebraic graph theory with other related knowledge,a control protocol is designed for the system,under which the system can finally converge to the group consensus of any non-zero projection parameters,and avoid Zeno behavior.The designed control strategy can also effectively reduce the loss of resources in the system and improve the speed of uniform convergence.The third part introduces the leader agent into the model of the second part,Therefore,the research results can be applied to some specific projects.A control strategy combining impulsive control and event-triggered control is designed for the leader-following group consensus problem in general linear multi-agent system.The results of this study further popularize the traditional leader-following consensus problem.A sufficient condition for the leader-following group consensus of any projection parameter is given in the results of this study by using the stability theory of impulse differential equations,matrix theory and inequality theory so that the traditional leader-following consensus problem will be further extended while avoiding Zeno behavior.In addition,an improved event triggering function is also proposed to avoid continuous communication between agents.The fourth part considers the change of the actual network environment,The consensus problem of general linear time-varying multi-agent systems with time-varying topologies is studied,A control protocol,which will greatly save the communication and control cost of the network to achieve consensus,based on event-triggered impulsive control is proposed in the fourth part,aiming at the consensus problem of multi-agent systems in the time-varying network connection environment.By using algebraic graph theory,decoupling method and impulse differential equations,the sufficient conditions for the consensus of multi-agent systems with time-varying characteristics under the event-triggered impulsive control are analyzed and derived.At the same time,it is proved theoretically that there is no Zeno phenomenon in the time series triggered by events.The correlative inferences of the system under the time-invariant topology are also given.Finally,a simulation example shows the validity of the theoretical results.