Distributed Bipartite Consensus Control For Multi-agent Systems Under Fixed And Switching Topology

As the research on computer and artificial intelligence technology continues to intensify,the multi-agent systems are widely used in biological systems,sensor networks,spacecraft formation flights and other fields because of the excellent performance in many aspects of flexibility and efficiency.However,the communication relationship between agents can be changed at any time once the multi-agent system has external disturbances,communication limitations or unknown obstacle.At the same time,for a long time research scholars have assumed that the agents achieve information interaction through cooperative cooperation,which does not match the multiple agents’ competitive relationship that exist in practical applications.Therefore,it is a topic of great research importance on how to guarantee system stability in the presence of confrontational systems and to investigate the bipartite consensus tracking of agents in different communication topology.In this thesis,we combine Lyapunov stability theory and structural balance theory to study system with simultaneous cooperative-competitive relationships,and construct a multi-agent model at the time of appearance of actuator faults and communication restrictions,and the appropriate conditions to achieve bipartite consensus tracking are proposed in different communication topology,respectively.Then,the bipartite consensus theory is extended to the case where there are multiple leader agents,and the study of bipartite consensus control theory is enriched by the design of containment control protocols based on observer estimation information to obtain sufficient conditions for the system to achieve containment control.The main research content and innovations in the full thesis are as follows.1)A linear multi-agent system with simultaneous relations,assuming a fixed and undirected communication topology is studied to achieve a bipartite consensus tracking problem in case of actuator faults.A hypothesis is given that the communication topology between the agents is both competitive and cooperative.Based on the symbolic characteristics,a state observer and a bipartite consensus control protocol are designed to give and prove the conditions for achieving bipartite consensus tracking under the premise of structural balance.Finally,the validity of the research methodology is demonstrated by numerical results and simulation plots for the relevant cases.The innovative point is to consider the stability conditions in the case of actuator faults in the follower agent system,and to study the bipartite consensus tracking problem of multi-agent system with simultaneous competitive relationship,based on the traditional consensus analysis which only assumes the existence of cooperative relationship between agents.2)An agent model with Markov switching topology is investigated to achieve bipartite consensus tracking under competitive relationships.In order to address the issue that the multi-agent communication topology can change at any time,the time-varying switching process of the agents’ communication topology is constructed as a Markov model.In addition,considering the effect of actuator faults of the follower agents,an extended state observer is designed to achieve an accurate estimation of actuator faults and the state of the agents,and an infinitesimal operator is used to verify the effect of the system on error convergence in a bipartite consensus control protocol,and a simulation example is used to demonstrate that the control protocol can achieve consistent tracking while using adaptive control to compensate for the effect of actuator failure.The innovation lies in choosing a more investigative Markov switching topology to extend the findings under a fixed topology to the case of stochastic switching of the communication topology,enriching the study of bipartite control under competitive relationships.3)To solve the problem of distributed bipartite consensus control under structural balance.The fixed topology is studied as a special case of Markov switching topology for systems where multiple leader agents exist.Next,a distributed state observer with relative neighbor information and a bipartite containment control protocol are designed by using structural balance theory.The criterion in terms of linear matrix inequalities that enables the system to ensure bipartite containment control in both cases is given,and the applicability of the designed control algorithm is also illustrated using simulation examples.The main innovation of the study is the problem of bipartite control in the co-existence of competitive and cooperative states,and the assumption that the communication topological states switch in time and that the switching process can be described and portrayed using a Markov model.