Back-propagation neural networks for nonlinear self-tuning adaptive control

A back-propagation neural network is applied to a nonlinear self-tuning tracking problem. Traditional self-tuning adaptive control techniques can only deal with linear systems or some special nonlinear systems. The emerging back-propagation neural networks have the capability to learn arbitrary nonlinearity and show great potential for adaptive control applications. A scheme for combining back-propagation neural networks with self-tuning adaptive control techniques is proposed, and the control mechanism is analyzed. Simulation results show that the new self-tuning scheme can deal with a large unknown nonlinearity.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>