Identification of a PUMA-560 two-link robot using a stable neural network

Abstract

A training procedure for a class of neural networks that are asymptotically stable is presented. The training procedure is a gradient method which adapts the interconnection weights, as well as the relaxation constants and the slopes of the activation functions used, so that the error between the expected and obtained responses is minimized. A method for assuring that stability is maintained throughout the training procedure is given. Such a network is used to identify a simulated nonlinear system and a PUMA-560 two-link robot.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>