SYSTEM IDENTIFICATION AND CONTROL USING NEURAL NETWORK

Abstract

System identification and control has always been a topic of interest and research for the control engineers. In spite of the large variety of methods available till date still the problem of identification of control of dynamic plants is not yet fully solved. In this dissertation identification and control of dynamic plants has been studied and implemented using neural network. The neural network used is a fuzzy 2 neural network which uses error back propagation algorithm for training. Fuzzy rules help us to model the uncertainties of the real world problem and neural network provides with very accurate and fast learning. Thus combining two results in a development of a very accurate and fault tolerant scheme for identification and control. Simulation studies are carried out on plants taken from literature and a comparison of performance with_existing methods is given.