Using a context-sensitive learning network for robot arm control

Abstract

In this paper, we propose a clsss of networks called contezt sensitive learning networks for use in the learning of complex nonlinear mappings. In particular, we present a network architecture for learning to control a robot arm by learning independently the different entries of the inverse Jacobian matrix. Computer simulation results showed that the network was able to learn the inverse Jacobian of the PUMA 560 arm for inverse kinematic control. The network also generalised reasonably well when unseen testing examples were presented to the network.