On Hybrid Adaptive And Pinning Consensus For Multi-agent Networks

As the basis of multi-agent network coordination control,consensus issues have been widely used in formation control of multi-robots,unmanned aerial vehicles,unmanned vehicle,flocking control,aggregation and synchronization problems and so on.In addition,information fusion between sensors such as distributed cluster and tracking can also be considered as the application of multi-agent consensus coordination problem,with the rise of sensor network.For better understand the complex network,explore the rules and provide valuable reference for practical application.It is necessary to study the consensus of multi-agent network.This thesis bases on the existing research results,using the algebraic graph theory,control theory and stability theory,the consensus problems of continuous-time and discrete-time multi-agent networks based on self-adaptive and constraining strategies are discussed separately,and a comprehensive containment strategy is given.The main contents of the dissertations are summarized as follows:Based on the existing research results,the multi-agent network topology is modeled by using graph theory.Combining matrix theory,stability analysis theory,and the related properties of M-matrix,the hybrid adaptive control and containment control are used to solve the consensus problem of continuous-time and discrete-time multi-agent networks.The pinning consensus problem of continuous-time multi-agent networks with general connected topology is discussed.Based on matrix theory and stability theory,the conditions for asymptotically achieving a uniform system are given.At the same time,a new novel control strategy is proposed according to the properties of M matrix.The strategy indicates that selecting a node with a zero degree as the object to be controlled can ensure that the system achieves consensus asymptotically,and a node with a out-degree greater than the in-degree can be selected in preference to other nodes to improve the system's convergence speed.In addition,the study indicates that the more nodes selected for containment,the faster the network convergence speed is not always true.In this thesis,we discuss the problem of pinning consensus of discrete time multi-agent system with general connected topology,and the algebraic condition criterion is given.The results show that the control gain and the coupling weight are related to the consensus achieving of pinning control for discrete system.Finally,the work of thesis is summarized,and the follow-up research work of the multi-agent network consensus problem is prospected.