Dynamic System Identification Using Time-Delay Feedforward Neural Networks: Application to Dc Motor

Abstract

The universal function approximation capabilities of multilayer feedforward neural networks make it a popular choice for modeling dynamic systems. In this paper, identification of dynamic system using time-delay feedforward neural networks with application to DC motor as a case study has been developed. The developed neural network model is a three-layer network with nonlinear (sigmoid) activation functions in the hidden layer and linear output layer with input-output delays. Simulation results showed that the neural networks are promising tool for dynamic system identification.