Distributed Coordination Control Of Linear Multi-Agent Systems With External Disturbances

Recent years,fruitful results on the distributed coordination control of multi-agent system have been carried out by many scholars.Each agent in the multi-agent systems communicates with their neighbours to plan their own behaviors such that the whole systems approaches a common value.From the aspect of group behavior,the main research contents of multi-agent systems coordination control include leader-follower tracking,consensus,formaiton and flocking,where the leader-follower coordination tracking and the consensus gain much more attentions.When there exists disturbance in each agent'dynamic,the existing coordination control laws won't work anymore.Hence,in this thesis,the leader-follower tracking and consensus problem of linear multi-agent systems with external disturbances is studied.The main contents and contributions of this dissertation can be summarized as follows:First,for the situation that only a part of the followers can obtain the information of the leader,by setting distributed tracking errors term,and using theory analysis method based on Lyapunov stability,we make the tracking errors converge to zero.To reject the external disturbances in the multi-systems,we set up a distributed disturbance observer which can suppress the disturbances by the feedforwad control channel.Since the Laplacian matrix of the directed graph is nonsymmetric,a novel Lyapunov function based on the relative state information is designed.By using the stability analysis,a sufficient condition of each agent's heterogeneous gain can be achieved.In the distributed coordinated tracking problem,the leader's state information can be used as the reference signal to design the control algorithm.However,for the consensus problem,there isn't any leader in the system,only the relative information of each agent's neighbours can be used.Thus a novel consenus algorithm based on the compound observer is proposed,where the state estimates are used in the feedback and the estimates of the disturbances are used in the feedforward.However,the disturbacne observer is denpended on the state observer.To solve this problem,a improved consensus algorithm which is only based on the disturbance observr is proposed.For the case that the global information is required for heterogeneous control gain,a fully distributed consensus control algorithm with adaptive control gains is designed in the directed graph.Based on the previous disturbance rejection consensus algorithm,the adaptive control gain is introduced using only the relative information.This kind of adaptive control gains varies with time and eventually converges to the desired control gains which can make the system to achieve the control objective.