Consensus Of Multi-agent Systems With Interference

With the development of science and technology,multi-agent systems(M ASs)have been widely used in recent years.However,in practical application,information transmission between individuals is often disturbed by some factors in the environment,such as the inevitability of noise and the timeliness of information exchange.At the same time,due to the constant movement of the agent,its connections will undergo some changes.This has stimulated interest in how topology changes affect the performance of the entire system.Based on the above,this paper gives the following three research results for M ASs:First,for the single integrator system,a nonlinear controller is designed under the condition of multiplicative noise interference.The controller contains linear coupling term to ensure the consensus of the M ASs.Based on the finite-time control technology,a nonlinear coupling term with a signed function is proposed to achieve finite-time consensus.In addition,the last term of the protocol is used to simulate measurement noise.It is assumed that the topology graph is undirected connected graph and the noise intensity matrix is symmetric.Finally,the results are extended to the case of topology switching.By using stochastic finite time stability theorem,Lyapunov function and It(?) formula,a sufficient condition for global stochastic finite time stability of the system is obtained.In order to satisfy the conclusion of the theorem,some parameters in the controller need to satisfy certain conditions.In addition,the effects of the parameters in the protocol and the algebraic connectivity of the graph on the convergence rate are also considered.Through simulation comparison,it is found that both small value of p and large value of ? can accelerate the convergence speed of the system.Second,the finite-time consensus problem of event triggered M ASs by multiplicative noise is studied.Design a nonlinear event triggered agreement,and consider a scalar agent.And in order to reduce the frequency control driver update,for each agent,assuming that the controller in only a few discrete time updated,these discrete time is determined by the proposed event trigger condition,and each agent has its own update time,and is only related to its neighbor state.Similarly,to achieve consensus,the control parameters need to meet certain conditions.The results are extended to the switching topology.Appropriate parameters are given in simulation to accelerate the convergence speed of the system.Thirdly,by using stochastic analysis,It(?) formula and some special inequalities,the pth moment consensus of M ASs which is affected by measurement noise and communication delay is obtained.The leader is considered in the system,and the leader is required to be globally reachable.A consensus protocol with time-varying gain is designed to deal with the effect of additive noise,and the gain is required to meet the convergence and robustness conditions.In addition,the time-varying delay is considered here,and a large upper bound on the time-varying delay is obtained by introducing some relaxation constant parameters.