IDENTIFICATION OF DYNAMICAL SYSTEMS USING NEURAL NETWORK

Abstract

Many studies have been undertaken in order to apply both the flexibility, learning and modeling ability of neural network, in the field of control system. We discuss methods for identification of dynamical systems using back-propagation algorithm. Basic concept related to identification and Neural network are discussed. We also consider the various methods available for identification of systems. Identification of dynamic system is explored via computer simulation using C and C++ languages. Simulation results shows that the- trained neural network model behaviour for the same input for which it is trained is perfect, but for different input, it tries to approximate the . behaviour of actual system.