Research On Global Consensus Control Algorithms For Multi-agent Systems Based On Approximator

In the past two decades,the problem of consensus control for uncertain multi-agent systems has become a hot issue in the field of control science.Especially,distributed control based on approximator has been widely concerned by scholars,and a large number of theoretical research results have been obtained.However,there are still some open problems worthy of further study and discussion.In this dissertation,the problem of global consensus control is studied.The research idea comes from the global stability of a single system.The research idea comes from the global stability of a single system.The universal approximator is used as a feedforward compensator to describe the uncertain nonlinear dynamics,and the leader signal is used as the input signal to approximate the uncertain nonlinear dynamics.Compared with the research method of global stability theory of single system,the research of global consensus theory of multi-agent system is more complex,and it is not a simple combination of multiple single systems.The research of multi-agent systems should consider not only the physical characteristics of agents themselves,but also the coupling characteristics between agents.In the multi-agent systems,the communication between agents is local.Each agent can only get the signals of itself and its neighbors,not every follower can get the signal of leader.Therefore,research on global consensus of multi-agent systems based on universal approximator,firstly,according to the network topology structure,the followers who can directly obtain the leader signal are classified into one group,and those who can not obtain the leader signal are classified into another category;then,different distributed control protocols are designed for different types of followers.Based on this design idea,this dissertation analyzes the global consensus of the following two kinds of uncertain nonlinear multi-agent systems and designs the controller.1.For the first-order and second-order uncertain multi-agent systems,under the undirected topology,distributed control protocols are designed respectively,and the quadratic Lyapunov function is constructed to analyze the designed distributed control protocol.A sufficient condition for the global consensus of the closed-loop system is derived.2.Aiming at the uncertain multi-agent systems with periodic disturbances,under the undirected topology,a distributed control protocol with repeated learning is designed,and a Lyapunov-Krasovskii functional is constructed to analyze the stability of the closed-loop system and sufficient conditions are given for the global consensus of the closed-loop system.3.For the first-order and second-order uncertain multi-agent systems with stochastic disturbances,under an undirected topology,based on the stochastic Lyapunov stability theory,a sufficient condition for the asymptotic consensus of the closed-loop system in the sense of probability is given.4.For mixed-order uncertain multi-agent systems with unknown directions,by adding virtual node states,an equivalent uncertain multi-agent systems of the same order is established.Under the undirected topology,a mixed control protocol is designed,gives sufficient conditions for the global consensus of the closed-loop system.5.For the mixed-order uncertain multi-agent systems with stochastic disturbances,the orderincreasing method is used to establish the equivalent uncertain multi-agent systems of the same-order.Under the undirected topology,based on the Lyapunov stability theory,Lyapunov function is constructed to prove the sufficient conditions for the global asymptotic consensus of the closed-loop system in the mean square sense.Finally,through simulation examples,the above theoretical results are verified.