On Anti-saturation Consensus Control Of Nonlinear Multi-agent Systems

Multi-agent systems(MASs)are widely used in many fields such as industry,military,medical treatment and transportation.In practical applications,MASs often possess nonlinear dynamics,and suffer saturation constrants due to the limitation of space,energy and actuator and sensor structures.Therefore,research on anti-saturation consensus control of nonlinear MASs has important theoretical significance and potential application value.In view of this,this dissertation focuses on the consensus control problem of nonlinear MASs with saturation constrants,by employing several advanced control theories such as impulsive control,fuzzy control and neural network theories.The main contents are as follows:An impulsive control strategy is proposed to solve the anti-saturation consensus problem for a class of nonlinear MASs with actuator saturation in the leader-following communication mode.By establishing the impulsive differential equation model for the errors between leader and followers,based on the convex hull property,the invariant set conditions are obtained to guarantee the consensus of the system,which ensures that the proposed impulsive control strategy can achieve leader-following consensus.The invariant set condition is further transformed into an optimization problem,and the impulsive control gain can be obtained while estimating the size of the attraction domain.The effectiveness of the designed control strategy is verified by the simulation results.A distributed impulsive control strategy is designed for the anti-saturation dynamic consensus problem of a class of nonlinear MASs with actuator saturation in the leaderless communication mode.The control strategy only collects neighbor information and implements controls at the impulsive instants,which has the advantage of reducing communication energy consumption and controlling energy consumption.By employing the time-varying Lyapunov function method,the information of MASs at impulsive instants is fully utilized.Therefore,the set invariance condition at a low conservativeness is obtained,and the anti-saturation dynamical consensus is achieved.This method can be applied to both undirected and directed graphs.The simulation results verify the effectiveness of the designed control strategy.To solve the anti-saturation consensus problem for a class of nonlinear MASs with both actuator saturation and time-varying delay,a distributed impulsive control strategy is designed in the leaderless communication mode.By means of the comparison principle of impulsive differential equations,the time-varying delay problem in nonlinear terms is solved,and the delay-independent sufficient condition is established to achieve dynamic consensus.The effectiveness of the designed control strategy is verified by the simulation study of a group of lead-acid battery systems.For a class of nonlinear MASs with actuator saturation,a distributed cooperative control strategy is designed based on the fuzzy model method in the leader-following communication mode.Based on the T-S fuzzy model,the error model is established for the errors between the time-varying leader and the followers with actuator saturation.The invariant set conditions in the form of linear matrix inequalities are obtained by means of convex hull properties.The estimation problem of the size of the attraction domain is further transformed into an optimization problem,where the control gain to be designed is regarded as a decision variable in the optimization process to decide the control gain and to estimate the attraction domain imultaneously.The control gain is thus obtained to achieve the attraction domain as large as possible.Finally,the effectiveness of the theoretical result is verified by simulation example.A distributed event-triggered consensus controller is designed based on neural network to achieve the anti-saturation practical consensus of MAS with sensor saturation and unknown nonlinearity in the leaderless communication mode.The designed controller only need to access neighbor information and update its own neural network weight at the triggering instants,which has the advantage of low communication load.Based on the hybrid system theory and comparison principle,the lower bound of the event-triggered intervals is obtained,and the adaptive updating strategy of the self-triggered neural network weight is established,which can ensure the anti-saturation practical consensus of the MAS.Numerical simulation of the MAS verifies the effectiveness of the designed control strategy.