A modified 3-layer perceptron for control of robot manipulators

Abstract

A modified three-layer-perceptron neural network model is developed for mapping and classification of spatiotemporal patterns. First a formal definition of an intelligent system and intelligent control scheme for robot control is presented, and the neural network approach is reviewed. Most of the neural networks to date have been very successful in pattern classification for static problems. However, the class of network developed is for control problems. Two simple simulations demonstrate the performance of the proposed network as a temporal associative network.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>