Group Output Consensus For Multi-Agent Systems With Reference Inputs

With the rapid development of networked control technology,the research on consensus of multi-agent system has been paid more and more attention,and the theory of consensus control of multi-agent system has been applied to different scientific fields,such as multi-robot cooperative work,multi-UAV tracking flight and multi-wireless sensor information fusion.As an important problem of consensus control,group output consensus include traditional global output,through the analysis of intelligent body physical model,the rational design of distributed control protocol,makes all the agent in accordance with the established group rules,and let each group of the agent's output quantity to achieve consistently,and each group's values to achieve consistently each are not identical.Based on algebraic graph theory and matrix analysis,Schur stability theorem and network prediction method,with communication constraints(time delays and data loss)grouped output consensus analysis of multi-agent system,design the consensus control protocol packet output,at the same time to realize the group with the reference input output system is necessary and sufficient conditions for consensus,and the group output consensus problems with leaderfollow mechanism and no leadership group output consensus were studied.The main research contents of this thesis include the following aspects:First,the stability and group output consensus of a class of high-order discrete heterogeneous multi-agent systems with communication constraints are studied in this thesis.A network predictive control scheme is proposed to use the information predicted state affected by communication constraints to actively compensate for communication constraints.The introduction of cooperationcompetition interaction eliminates the constraints of the in-degree balance and relaxes the constraints of the topology.For high-order discrete heterogeneous systems,a dynamic output feedback distributed control protocol is designed based on predictive information.Combining consensus analysis and stability analysis,the sufficient and necessary conditions for the stability of the two groups with dynamic output feedback and the consensus of grouped output are derived,and the conclusion is extended to multiple groups.Secondly,considering that the special agent in the multi-agent system acts as a leader to track the externally given reference input,leading the entire system to converge to the reference input.In this thesis,the leader-following mechanism is added to the research of group output consensus,and the problem of group consensus tracking in multi-agent systems is studied.When a leader in a multiagent system receives an external reference signal,and the agent designs a distributed control protocol based on neighbor node information.The necessary and sufficient conditions for consensus tracking of two groups of grouped outputs with reference inputs are derived.Through the analysis of the two groups of multi-agent systems,this thesis extends the necessary and sufficient conditions for consensus tracking of grouped outputs to multiple groups.