Research On Cooperative Control And Output Regulation Of Multi-agent Systems

In the recent years,many scholars have paid attention to cooperative control of multi-agent systems.Multi-agent systems could complete a lot of tasks which the single agent can not complete,and it has extensive application in society,industry,national defense and so on.Cooperative control of multi-agent systems contains the filed of consensus,cluster,formation,tracking,and one of the most important issue is output regulation of multi-agent systems.If the output of the exosystem can be tracked by the outputs of the other agents(or/and the disturbance be rejected)based on the appropriate distributed control law,these problem is called output regulation problem.However,due to the complicated structure of agents and ex-osystem,output regulation has been become a challenging issue in control theorem.This dissertation considers continuous and discrete time multi-agent systems,and studies the cooperative control and output regulation of multi-agent systems via distributed dynamic state and dynamic output feedback control law based on graph theory,matrix analysis,stability and differential equation theory.Some numerical simulation are used to illustrate the validity of the theoretical results.The main contents are listed as follows:(1)A general global method has been designed via distributed dynamic state feedback control law to solve output regulation problem of state-coupled linear multi-agent systems.The transformation of the Jordan canonical form was used to stabilize the closed-loop control without consider the disturbance,such that the stability of the whole closed-loop system was equivalent to the stability of multiple sub-systems.Sylvester equation and internal model theory were adopted to achieve the objectives of robust output regulation for every initial condition in the state space.(2)Cooperative output regulation for heterogeneous multi-agent systems with par-tial unmeasurable agents' states has been considered.All the agents except exosystem can be classified into two groups.Agents in the first group can be measured by themselves and their neighbours.State variables are not fully accessible for direct communication and full order Luenberger observers were constructed for the unmeasurable agents.We gave a state feedback control law to solve the output regulation problem under the communication topologies based on both measurable and unmeasurable agents.A distributed compen-sator was used and dynamic state and output feedback control laws have been designed in solving the robust output regulation of uncertain multi-agent sys-tems.(3)Distributed tracking control of leader-follower multi-agent systems with direct-ed communication topologies has been solved.The directed communication graph need to contain a spanning tree.It was supposed that the leader has the different internal dynamic with the followers,and each agent can not mea-sure the states' information of itself and its neighbors directly.Therefore,the state information of the agents can be divided into the measurable and unmeasurable parts,and a reduced-order observer was designed for the first time for the multi-agent system which contains the external disturbance by using an appropriate transformation.At last,the result was extended to the continuous-time multi-agent systems.(4)Most existing results usually depend on the distribution of the eigenvalues of Laplacian matrix in solving the cooperative control of discrete-time multi-agent systems.In order to handle this problem,we designed a stabilized-region regulator method.A distributed dynamic feedback control law was designed such that the distribution of the eigenvalues of Laplacian matrix could be regulated into the specified region by choosing the appropriate parameters,then some sufficient conditions used to guarantee output synchronization are presented.(5)Output regulation problem of heterogeneous discrete-time multi-agent systems with multiple leaders subject to structural uncertainties under the directed communicating topology has been solved.The information exchange channels of the followers do not need to be connected.A novel distributed dynamic feedback control law based on internal model compensator was developed to achieve the robust output regulation problem of multi-agent systems with multiple leaders,where the convex hull information of the reference outputs is embedded into the internal model compensator.