Real-Time Control of Robot Manipulators by Neural Networks

Abstract

This paper describes an innovative approach for real-time kinematic control of robot manipulators by using neural networks. While different approaches in this field have appeared in the literature, this paper shows how control theory can be incorporated into the neural network design so that the quality of the solution can greatly be enhanced and the feasibility for real-time implementation can be achieved. Extensive simulations illustrate the attractive features of this proposed approach.