Research On Consencus Of Multi-agent System Based On Event-trigger

Since multi-agent systems have broad application prospects and event-triggered control has the advantages of saving communication resources,it is easy and meaningful to study the consensus of MAS based on event-triggered control.The main research content mainly includes the following aspects:Firstly,this paper considers that when the communication topology is an undirected connected graph,studies the event-trigger consensus problem of a linear multi-agent system with a first-order integrator model.In order to reduce the problem of continuous communication between agents,a distributed event trigger algorithm is adopted for each agent.In this algorithm,the data update of each agent's controller is driven by events.A given trigger condition determines whether to update the controller and transmits information to neighbor agents.The controller is not updated in the interval between two events.This can greatly reduce the number of communications and save resources compared with the previous continuous communication method between agents.In addition,this article discusses a multi-agent system event-trigger consensus algorithm based on periodic sampling control.By introducing a periodic sampling scheme,Zeno behavior can be avoided naturally.Since communication can only be carried out at the sampling moment,and the control input of each agent remains unchanged during the event time interval,continuous communication with neighboring agents is no longer required,which effectively reduces the number of control input updates.The Lyapunov function and the principle of Lasalle's invariant set are used to prove that the system is stable,and can achieve consistent under the given event-triggered control algorithm,and Zeno behavior will not occur.Secondly,considering the event-triggered consensus problem of continuous time MAS with general linear dynamics.When the communication topology is an undirected graph,a distributed event-triggered predictive state feedback consensus control algorithm is adopted for each agent.Under this algorithm,there is no need for continuous communication between the agents.The information transmitted by the neighbor agent is determined by the trigger condition.It is further analyzed that the system is stable under this algorithm,can achieve consistent state,and does not produce Zeno behavior.Next,an observer-based event-triggered predictive state feedback consensus controller under a directed graph is studied,which is used to solve the problem that the state of the agent is difficult to measure in actual situations,and the relevant analysis is given to illustrate the effectiveness of the algorithm.Finally,considering the general linear dynamic multi-agent system,the event-triggered consensus problem with external disturbances.Under the undirected communication topology diagram,in order to suppress disturbances,a distributed event-triggered H? consensus control algorithm is adopted for each agent.Under this algorithm,there is no need for continuous communication between the agents,and it is determined whether the agents need to transmit information and update the controller through a given event-trigger condition.It is further analyzed that the system is stable under this algorithm,can achieve consistent state,and does not produce Zeno behavior.For all the algorithms in this paper,numerical simulations are given to prove that the simulation results are true and reliable.