Output Feedback Of The Nonlinear System Based On Adaptive Neural Network

For the nonlinear non-minimum phase systems with uncertainty, its output feedback stabilization problem has become a hot issue of nonlinear research and has certain value for study. The neural network has strong adaptive complex en-vironment and self-learning ability of multi-objective control requirements, and can approximate any nonlinear continuous functions by arbitrary precision. For exist a uncertainty of robust control systems, through design and robust controller, the closed-loop system can remain stable at the same time and guarantee the dy-namic performance of quality. This article integrates neural network and robust control advantages, and resolve the output feedback control of the nonlinear sys-tem and the nonminimum phase system.First of all, the paper introduce and analysis the Back-stepping output feed-back control method, through actual simulation, pointing out the controller defi-ciency. Based on this, the paper based on neural network and robust control ad-vantages, puts forward a new output feedback controller. Firstly, in this paper system has been Taylor expansion, then introduced two uncertainties, namely: matching uncertainties and zero dynamic modeling error. It makes the zero dy-namics modeling error to satisfy uniform boundedness using the assume condi-tion, and using MLP neural network nonlinear mapping capability to make that approached the matching uncertainties; The observer of closed-loop system that its output can be measured is designed. In the paper, simulation results of the TORA system show that this control algorithm is effective, and the convergence rate of the corner achieve steady state in short time. Convergence rate of the dis-placement was Obviously improved.Improved by the introduction of the adaptive robust control items, and it is a control method to approximate the error of the neural network by the maximum extent . According to the simulation result, by the improved controller compares with the original Back-stepping output feedback controller, it is slightly faster in the response speed of the corner. In the other control performance, the improved controller than without improved control performance is better.Finally, in this paper, the stability of the closed-loop is proposed with output feedback controller based on adaptive neural network, and it gives the stability proof conclusion.