Research On Tracking Control Based On Consensus Of Multi-Agent Systems

In recent years,the theoretical results of distributed consensus control of multi-agent systems are widely used in engineering fields,such as aircraft formation,sensor network configuration,cooperative control,and target tracking,which displays great application prospects and important academic research value.The design of information exchange rules is the core of the consensus problem solving,which purpose is to make all individual states of the agent network converge to a certain expected value.In this paper,we mainly study the problem of target tracking derived from consensus control,and and expatiates on the relationship between consensus of multi-agent systems and tracking control.The multi-agent systems are divided into two parts: the leader and the follower.The communication mode among them is described by an algebraic graph.This paper introduces the necessary theorems and stability theories in the proofs are given.The main research content of this article is as followsFirstly,the tracking control problem of multi-agent systems with the additive stochastic noises is studied under directed graph.When vertex of the leader is globally reachable on the digraph,the proposed fully distributed consensus control protocol,where adaptive rate of the consensus gain only depends on the tracking error for each agent,by using algebraic graph theory,adaptive control theory and stochastic theory tools,two conclusions are obtained:1.Each follower can track the leader in mean square;2.The consensus gain of each agent tends to the optimal value.Finally,the effectiveness of the proposed algorithm are validated by two simulation examples.Secondly,the tracking control problem of leader-follower multi-agent systems with measurement noises and time-delay is studied under directed graph,where acceleration of a active leader is unknown to the followers in directed topology.Each agent only shares the relative position of its neighbors.A distributed consensus controller is proposed and divided into two components: one is the neighbor-based velocity estimator to the leader for the followers,the other is the neighbor-based feedback control law for the followers.In particular,the effect of decomposition errors to the time-delay is analyzed.By using algebraic graph theory and stochastic analysis theory,it is theoretically proved that the consensus tracking algorithms can guarantee tracking errors converge to zero in mean square.Two simulation examples are provided to validate the effectiveness of the proposed algorithm.At last,the finite-time consensus tracking control problem of second-order multi-agent systems with stochastic noise is studied under undirected graph.The agent acquires neighbors' measurement information with stochastic noises.Selecting appropriate Lyapunov functions for errors differential equations of stochastic multi-agent systems,and construct some suitable intermediate variables,to process the Hessian matrix generated during the derivation by using of the block matrix method.The tracking control protocol designed based on combination with graph theory and finite-time stability theorem ensures that followers can track dynamic leaders in limited time.The simulation example verifies the correctness of the finite time tracking protocol.