Research On The Consensus Of Multi-agent System And Its Application In Rendezvous Control

In recent years,with the progress of industrial technology and the development of distributed control theory,more and more distributed intelligent devices are put into solving the complex scene of global task scheduling and coordination,such as UAV formation,robot search and rescue,etc.In order to cooperate with each other to accomplish a specific task,these intelligent devices often need to construct a network topology for information interaction,and adopt a specific distributed control algorithm to make decisions and control behaviors.Therefore,the theory of multi-agent can be used to model this kind of distributed intelligent system,and applied to simulation or actual task execution environment.In the field of multi-agent system,the consensus protocol is the key research content.The solution of this problem implies that the state values of all agents in the system converge to the same value,so that the subsequent distributed tasks can be performed more stably.The state value of agent can refer to speed,position,viewpoint,etc.When the goal is to make the positions of agents reach a consensus,we call it the rendezvous problem.How to design a high-performance distributed rendezvous control algorithm to make all the agents asymptotically converge to the same value in finite time is the difficulty of research.The research contents and achievements of this paper are as follows.In this paper,the rendezvous problem of multi-agent system under normal condition and communication failure condition is studied respectively,and the corresponding rendezvous control algorithms are proposed to achieve the system consensus.On the basis of classic algorithms,several new algorithms improve the calculation method of candidate points according to the neighborhood information,and explore the influence of relative neighbor graph as the topological connection structure on the system performance.In specific scenarios,these new algorithms can make the multi-agent system obtain better rendezvous performance.When the multi-agent system is in normal operation,two distributed algorithms are proposed to solve the multi-agent rendezvous problem with limited communication range in discrete time.Firstly,a convex hull combination algorithm(CHCA)is proposed.Each agent can calculate two kinds of candidate points according to the neighbors in the communication range,and combine with the constraint set to calculate the candidate points that are more conducive to the internal movement of the agent to the convex hull.Compared with the circumcenter algorithm,the convex hull combination algorithm can reduce the system rendezvous time under the initial low-density topology distribution.Moreover,by combining the relative neighbor graph as the constraint set with the CHCA,the new algorithm can reduce the rendezvous time of the system under the initial high-density topology distribution.Then,the sufficient conditions for the consensus of the two algorithms are proved.Finally,a large number of simulation experiments show that under the corresponding initial topology distribution,the two new algorithms can make the system rendezvous faster than the classic algorithm.In the case of communication failures,a limit convex hull algorithm(LCH)is proposed.Based on the traditional framework,the LCH constructs a minimum possible convex hull by combining the state information of the historical and current communication neighbors,and the agents gather to the interior according to the circumcenter of the convex hull as a candidate point.On the one hand,it increases the neighborhood reference range of the agents,on the other hand,it guarantees the convergence of the system.At the same time,the influence of the initial relative neighbor graph as the constraint set on the system rendezvous performance is studied.Then,the convergence analysis of the new algorithm under communication failures is proved by using the correlation theorem of the classic rendezvous framework.Finally,the simulation results show that compared with the classic algorithm,the new algorithm can reduce the rendezvous time and mobile distance when the communication failures are more frequent.