Research On The Fixed Time Consistency Of Nonlinear Multi-agent Systems

With the combination of information technology,communication technology,computer science and control technology,traditional industries are gradually becoming more and more automated.The research boom in the field of artificial intelligence has been set off,and the Internet of Things has become one of the vanes of modern development.The multi-agent system evolved from the cluster phenomenon of nature are increasingly of research value.Among them,the consensus problem of multi-agent systems has been widely concerned by scholars for a long time.Compared with the asymptotic consensus control protocols,the finite-time consensus control protocols have faster convergence speed and stronger robustness.Therefore,it has become a research hotspot in recent years.As an extension of finite-time consensus problem,there exists few research results of fixed-time consensus problem.The fixed-time consensus is that the multi-agent systems can reach consensus within the prescribed settling time regardless of the initial conditions.This thesis will mainly focus on the fixed-time consensus problem of nonlinear multi-agent systems.Firstly,the fixed-time consensus of first-order heterogeneous nonlinear multi-agent systems is investigated.Assuming that each agent of the multi-agent system has inherent heterogenous nonlinearity and the nonlinearity satisfies growth condition.In this case,dynamic signals are designed for each agent,and a new control protocol is designed to make the multi-agent system reach fixed-time consensus.The fixed-time leader-following consensus is further discussed.Considering there exists a leader in the first-order nonlinear multi-agent system,whose input information is unknown but bounded,with the aid of Lyapunov function method,the fixed-time leader-following consensus is achieved.Secondly,the fixed-time practical consensus of a class of more general first-order heterogenous nonlinear multi-agent systems is investigated.Assuming that each agent of the multi-agent system has completely unknown inherent heterogenous nonlinearity.The approximation of the nonlinearity can be achieved by using radial basis function neural network.Adaptive control law is designed to adjust the parameters.The corresponding control protocol is designed either.Then,the fixed-time leader-following practical consensus is further discussed.Considering there exists a leader with unknown bounded input,with algebraic graph theory and Lyapunov stability theory,it is proved that the followers can approach the small neighborhood centered of the leader under the designed control protocol.Finally,the fixed-time consensus of second-order heterogenous nonlinear multi-agent systems is investigated.Assuming that each agent with second-order dynamics has completely unknown inherent heterogenous nonlinearity.Using radial basis function neural network to approximate the nonlinearity.Recursive design method is applied to design the adaptive control law and control protocol.Then using sliding mode control to study the fixed-time leader-following consensus for second-order multi-agent systems.Assuming that the nonlinearity of each agent satisfies the growth condition and there exists a second-order leader with unknown bounded input.Observer system is designed for each agent to accurately estimate the statement of the leader.By choosing proper sliding surface,with the aid of algebraic graph theory,Lyapunov stability theory and fixed-time consensus theory,the control protocol is designed to make the followers track the leader within a fixed settling time.