Control Analysis Of Consensus Based On Multi-agent Systems

At present,the research on complex networks has matured,and then the studies on multi-agent systems in the complex networks are not only an important content,but also a significant part of distributed artificial intelligence control research.Moreover,cooperative control of multi-agent systems,among which consen sus of multi-agent systems is the fundamental problem,is one of hotspots in the field of control research in today's world.The consensus of multi-agent system represents t hat each agent in the system through information interaction and adjustment of proces sors will achieve the result of convergence.During the process of achieving consensus,consensus protocol is rule of information exchange and adjustment between the agen ts in the system,and the protocol of information exchange is crucial to achieve conver gence for the multi-agent system.Besides,the researches on the consensus of multi-agent systems have been used in many practical problems,such as UAV tracking,UAV formation control,multi-sensor information fusion,systems biology,physics,formation control and so on.Except that,event-triggered control and impulsive control has been proposed as an alternative to time-triggered control in recent years.A large number of research results show that event-triggered control is an effective way to reduce communication cost,and improve the quality of communication and network performance.The dissertation surveys the consensus problem of general linear multi-agent systems based on distributed event-triggered control and impulsive control.After systematically introducing of relevant research status,research work focuses on modeling and control in this paper,including:First of all,a simple introduction about some basic knowledge of complex networks is given,including the graph theory,the convex hull,impulsive control mechanism,and the event-trigger mechanism.At the same time,it contains consensus and containment consensus of cooperate behavior in the specific analysis of the paper.Then,given that in the process of the agent movement,it sometimes needs to achieve certain formation behavior between leaders and followers,we propose mathematical model of multi-agent system containment consensus based on the intermittent control strategy.Regardless of the angular velocity,namely considering the straight line direction movement,this article presents the intermittent sampling control protocol.Then under the control mechanism,we analyze how the multi-agent systems realize containment consensus,and obtain necessary and sufficient conditions of achieving containment consensus.Finally we demonstrate some numerical simulations to verify the effectiveness of the proposed algorithm.Afterwards,given fish swarm,we improve the model in the above and add to the state of the angular velocity,transferring the straight line movement into a circular motion.In order to analyze this problem better,we make a exchange between plane rectangular coordinate and polar coordinate,thus simplifying the model.In this part,we design impulsive containment control strategy,and obtain the necessary and sufficient conditions to realize the containment consensus.To closer to the actual situation,we also consider the communication between intelligent agents with constrained condition,and design the impulsive containment control strategy under random quantitative mechanism,thus analyzing how the multi-agent systems realize the movement in the circles.Finally,our theoretical analysis is verified by numerical simulations,which verify the effectiveness of the algorithm.And then,periodical event-triggered consensus with ideal relative state information is studied.Firstly,a distributed event-triggered controller is designed,so that the controller of each agent only triggers at its own time constants.What's more,the case when agents can only receive quantized relative state information is investigated.And numerical examples are provided to illustrate our theoretical results.Finally,the full dissertation is summarized and further research and development directions are proposed.