Neural networks in robotics: state of the art

Abstract

This paper provides a short review of the neural network approach to system control with reference to robotic systems. Starting with an exposition of the main neurocontrol architectures, the paper overviews the literature on the application of neural networks to robot kinematics, dynamics, path planning and motion control, including some work on neurofuzzy control. To appreciate better robot neurocontrol, an unsupervised robot neurocontroller is presented in some detail. The paper includes a discussion of the criticism made to the neural control paradigms, and an outline of some interesting areas for further research.