Title: A Morlet wavelet neural network-based online identification and control of coupled MIMO systems

Authors: Tarun Varshney; Satya Sheel

Addresses: Electrical Engineering Department, MotiLal Nehru National Institute of Technology, Allahabad, Uttar Pradesh, 211004, India ' Electrical Engineering Department, MotiLal Nehru National Institute of Technology, Allahabad, Uttar Pradesh, 211004, India

Abstract: A new approach has been developed for the control of a coupled, non-linear MIMO system. A quasi-diagonal wavelet neural network (QDWNN)-based online identification and control scheme for coupled, non-linear MIMO system is presented in this paper. The new algorithm is based on control error process and does not need accurate mathematical model of the system to be controlled. Firstly an online parameter-tuning algorithm for multivariable PID-controllers is described and then it has been tested for two model systems. Simulation results show that the proposed control scheme minimises mean square error (MSE) and mean absolute error (MAE) remarkably, with reduced control effort, without slowing down the rise time.

Keywords: quasi-diagonal wavelet neural networks; QDWNN; PID tuning; back propagation algorithms; BP algorithms; Morlet wavelet; online identification; online control; coupled MIMO systems; nonlinear MIMO systems; parameter tuning; multivariable control; simulation.

DOI: 10.1504/IJAAC.2012.051883

International Journal of Automation and Control, 2012 Vol.6 No.3/4, pp.246 - 260

Received: 04 May 2012
Accepted: 08 Sep 2012
Published online: 07 Nov 2014 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article