Adaptive Control Research On A Class Of Nonlinear System Based On Neural Network

Neural network adaptive control is designed by combining the neural network’scharacteristics with the adaptive control technology’s. It gave full play to the advantage ofboth and provides a new method to study the nonlinear system control.First this thesis gives a view of the recent studies on neural network-based adaptivecontrol systems. Focus on nonlinear identification and adaptive control with the way offeed forward neural network, especially deeply research the structure and algorithms of BPnetwork and RBF network. Based on neural network adaptive control process, the controlproblem of a class of nonlinear system is researched. Controller is designed arming at theunknown functions which used the BP neural network as approximation model existing innonlinear system. At last, the program is designed to test the effectiveness of the controller.The results of simulation by MATLAB verify the effectiveness of this adaptive controller.The main work of this paper is organized as follows: the method of neural networkadaptive control and the artificial neural network are introduced in detail from the firstchapter to the third chapter, which focuses on the feed forward network and the simulationon the approximation ability of the neural network; the forth chapter is devoted to theknowledge of system, and verify the rationality and effectiveness of the for nonlinearsystem identification by using of feed forward neural networks through simulation; thefifth chapter includes the simulation based on a class of nonlinear system instances for thecontroller designed and the analysis of the simulation results; in the end words theconclusion of this thesis is raised, and the development prospect of neural networkadaptive control is expected.