Research On The Distributed Bipartite Containment Control Of Heterogeneous Multi-agent Systems

In this paper,for heterogeneous multi-agent systems(MASs),considering the cooperation-competition communication network topology,the bipartite output containment control problem is studied.The dynamics and dimensions of agents’ states are different for heterogeneous MASs,which brings great challenges to design distributed control protocols.Considering the practical significance of the bipartite output containment control problem of heterogeneous MASs,this paper discusses the problem as follows:First,a state feedback control protocol is designed on the basis that states of agents are measurable.A fully distributed dynamic compensator is proposed to estimate the convex hull of leaders and its symmetric convex hull.The time-varying gain of the compensator is designed to not rely on eigenvalues of the Laplacian matrix of the communication graph,which makes the compensator not affected by the scale of the communication network.The feasibility of the compensator is demonstrated by establishing a Lyapunov function and based on La Salle’s invariance principle.Then a state feedback control protocol is proposed based on the compensator.The gain matrix of the control protocol is determined by the property of Hurwitz matrix and an output regulation equation to achieve the bipartite output containment control.Finally,a simulation is used to demonstrate that the proposed protocol is effective.Considering the situation that states of agents are difficult to measure,an output feedback control protocol further is designed.Different from the Luenburger observer in previous studies,a distributed observer using neighbor information is designed in order to estimate states of agents.Since the closed-loop error system matrix in heterogeneous MASs becomes complicated,an auxiliary system is constructed to analyze the stability of the error system.Then,an output feedback control protocol based on the distributed observer and the fully distributed dynamic compensator as in Chapter 3 is proposed to achieve the bipartite output containment control.Finally,a simulation compares the advantages of the distributed observer over the Luenburger observer,and verifies the validity of the proposed theory.