Adaptive Finite Time Intelligent Control For Several Classes Of Nonlinear Nonstrict-Feedback Systems

Since the 50 years in the 20 th century,research on control theory has received more attention than ever before.A large number of effective and practical control methods have been proposed by experts and scholars.The classical design methods based on the Lyapunov stability theory can realize asymptotic stability in which the state of the system converges to the origin when the time becomes infinite.However,in many practical applications,the target system is often expected to have the ideal effect of achieving control aim within a limited time.Therefore,whether it is based on the needs of theoretical development or the needs of practical application,the finite time control theory needs to be further reformation.Based on the above considerations,this thesis discusses some controller design problems for nonlinear nonstrict-feedback systems using nonlinear system theory and Backstepping control techniques,where our controllers stabilize the closed-loop system along with the good practical tracking performance.It has been demonstrated that excellent tracking performance is achieved.The summary of the main work is as follows:(1)For a class of single-input single-output(SISO)nonlinear systems,Chapter 2 designs a new control strategy based on quasi-fast finite-time stabilization criterion.Here,the ambiguity functions contained in the system are approximated using neural networks during controller design,and dynamic surface technology is used for the “explosion of complexity” problem that occurs during the controller design.The functions that have all states of the system are processed using the unique properties of the neural networks.Finally,through stability analysis,a numerical example and a practical example,the tracking error signal of the system can be converged to a adjustable small region of zero,and all states of the closed-loop system can eventually reach the ultimately bounded.(2)For a class of multi-input and multi-output(MIMO)nonlinear nonstrict-feedback systems with perturbations,we investigate an efficiency adaptive neural network tracking control scheme in Chapter 3.The superiority of this control strategy is that the scheme effectively solves the“singularity” issue that arises during controller design using the command filtering technique.The uncertain functions contained in the system are approximated using neural network during controller design;and then using the structural characteristics of RBF neural networks,the functions containing all state variables are simplified to reduce the design difficulty.The proposed control scheme can make the tracking error of the studied system reach the pre-defined performance requirements in a finite time,and also make the states of the closed-loop system eventually reach boundedness.Finally,an actual example is used to verify the effectiveness of the designed scheme.So far,research on the finite time control problem of nonlinear systems is still in the stage of continuous improvement.Compared with the existing results,the adaptive finite-time intelligent control problem for several classes of nonlinear systems with unknown functions is preliminarily studied in this thesis.However,there are still many related control problems need to be solved.For example,fast finite time control for nonlinear systems,finite time control for discrete-time nonlinear systems,fixed-time control for nonlinear systems,etc.