NEURAL NETWORK BASED ADAPTIVE CONTROL OF NONLINEAR SYSTEMS

Abstract

Multi layer neural networks are now widely used for nonlinear function estimations. This property can be used to estimate the states of a nonlinear system and approximate the system dynamics without using model or knowledge of the dynamics of the plant. In this paper, we implemented three layer, feedforward neural network to estimate the dynamics of a nonlinear system, inverted pendulum and use it as a ieedfurwad compensator. The stability and performance of the neural. network based adaptive controller is discussed and simulation results are given