Improved Adaptive Dynamic Surface Control For A Class Of Nonlinear Systems

This dissertation investigates the control problem of nonlinear system.Considering the relationship between system dynamic performance and control quality,system security under denial of service attack and system cooperative control,we research the improved adaptive dynamic surface control of a class of nonlinear systems by a series of theories such as dynamic surface control technology and adaptive control method.The main research work includes the following three aspects:Firstly,an adaptive neural network state feedback control strategy is designed for non-affine pure-feedback uncertain systems based on nonlinear gain and recursive sliding-mode dynamic surface.A kind of nonlinear gain function is introduced into the traditional framework of dynamic surface control(DSC)to make a compromise between control accuracy and transient performance.Recursive sliding-modes with the error between subsystems are constructed to expand the value range of adaptive parameters and filter parameters.Radial basis function(RBF)neural networks(NNs)are adopted to approximate unknown functions in process,and novel adaptive update laws are established to estimate the parameters of NNs approximation errors.To reduce the action number of the actuator,an extended control strategy in an event-triggered manner is proposed.It allows the actual control signal to be updated only when certain conditions are met.Considering the existence of unknown state variables in the system,an adaptive dynamic surface output feedback control strategy is proposed,which extends the application of nonlinear system control strategy in practical engineering.By the Lyapunov function,it is proven that the above control strategies can force the tracking error arbitrarily small and guarantee all the signals in the closed-loop system uniformly ultimately bounded.Finally,MATLAB simulation results are provided to verify the effectiveness of the proposed control strategy.Secondly,on the basis of the above research work,we address the secure control issue of non-affine nonlinear systems under denial of service(Do S)attacks.As for the situation that the system information cannot be measured in specific period due to the malicious Do S attacks,we design a neural networks state observer with switching gain to estimate internal states in real time.Considering the error and dynamic performance of each subsystem,we introduce the recursive sliding-mode dynamic surface method and a nonlinear gain function into the secure control strategy.The relationship between the frequency/duration of Do S attacks and the system's ability to withstand attacks is established by the proposed inequality condition.By the average dwell time(ADT)method and Lyapunov stability theory,the system stability and the boundedness of all closed-loop signals are proven.Finally,simulation results of flexible-joint manipulator and numerical example are presented to verify the effectiveness of the proposed secure control strategy.Thirdly,a communication event-trigger-based adaptive cooperative control strategy is proposed for consensus tracking of a class of uncertain nonlinear multi-agent systems(MASs)under directed graph.A distributed estimator based on neighbor's triggered output is designed to provide the leader's trajectory for parts of followers who are not able to access to leader's information directly.Recursive sliding-modes with the error between subsystems are constructed and nonlinear gain functions are applied to improve dynamic performance for individual follower.Also,a novel adaptive parameter is introduced to make the parameter of neural networks' weights dramatically reduced.In Lyapunov theory,it is proven that all signals of the closed-loop system are ultimately bounded with the consensus tracking error converging to a neighborhood around the origin and there exists non-Zeno behavior.Two simulation examples validate the effectiveness of the proposed strategy.