Study On Control Issues For Nonlinear Multi-agent Systems

With the rapid development of science and technology,control systems have become increasingly complex,the control system of today is not the simple,easily describe,low calculation cost single system into a complex,high cost multi-agent systems,which bring very great difficulty for system analysis.By adopting distributed control technology,the multi-agent systems can realize that single system cannot achieve the control objective,it is widely used in biological system,multi-robot network,formation control of uavs,network resource allocation control and other fields.The consensus\tracking control problem is the basic issue of multi-agent systems control research.The consensus control of multi-agent systems is that the states or outputs of agents can reach an agreement based on information communications amount agents.The tracking control of multi-agent systems is that all the outputs of followers converge to the small neighborhood of the leader output.On the other hand,complex transmission environment will result in the occurrences of numerous factors,such as actuator failure,actuator saturation,event-triggered,input quantization,antagonistic interactions and so on.Combined with the above problems,this paper conducted the following research:1)The problem of cooperative fault-tolerant control for stochastic nonlinear systems network with actuator failures and input saturation is studied.The fuzzy neural networks are employed to estimate the unknown functions and stochastic disturbance terms.To analyze the non-differential saturation nonlinearity,a smooth nonlinear function of the control input signal is constructed to estimate the saturation function.A novel adaptive fault-tolerant control protocol is proposed by using back-stepping design technique.By using the stochastic Lyapunov functional strategy,it is proved that all the followers outputs eventually converge to a small neighborhood of the leader output,and all the signals in the closed-loop systems are bounded in probability.Finally,the performance of the proposed control strategy is illustrated through simulation results.2)The problem of event-triggered fuzzy adaptive quantization control for random non-affine multi-agent systems is studied.The fuzzy logic system is used to estimate the stochastic disturbance term and unknown nonlinear functions.A nonlinearity decomposition method of asymmetric hysteresis quantizer is proposed by applying sector bound property.Moreover,to reduce the communication burden,an adaptive event-triggered protocol with a varying threshold is constructed.Based on the back-stepping technique and stochastic Lyapunov function method,a novel adaptive event-triggered fuzzy control protocol and adaptive laws are constructed.By using stochastic Lyapunov stability theory,it is demonstrated that all signals are bounded in the closed-loop systems in probability and all the outputs of followers converge to the neighborhood of the leader output.Simulation results illustrate the effectiveness of our proposed scheme.3)The problem of bipartite consensus tracking control for nonlinear network systems with antagonistic interactions and backlash-like hysteresis is studied.The generalized networked multi-agent systems model is considered in which every agent is an independent individual;and this model allows competitive and cooperative interactions to coexist.The Gaussian function is introduced to simulate cooperation and competition between two agents and radial basis function neural networks are applied to estimate an unknown nonlinear function.Based on back-stepping technology,we propose a new distributed adaptive neural control protocol that not only realizes bipartite consensus control but also ensures that the bipartite synchronization error converges to within a small range near the origin,and all signals are bounded in the closed-loop system.Finally,we present a simulation example to illustrate the obtained results effectiveness.