Learning convergence in the cerebellar model articulation controller

1 January 1992

journal article
Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Neural Networks

Vol. 3 (1) , 115-121
https://doi.org/10.1109/72.105424

Abstract

A new way to look at the learning algorithm in the cerebellar model articulation controller (CMAC) proposed by J.S. Albus (1975) is presented. A proof that the CMAC learning always converges with arbitrary accuracy on any set of training data is obtained. An alternative way to implement CMAC based on the insights obtained in the process is proposed. The scheme is tested with a computer simulation for learning the inverse dynamics of a two-link robot arm.

Keywords

This publication has 6 references indexed in Scilit:

CMAC: an associative neural network alternative to backpropagation
Proceedings of the IEEE, 1990
Generalizing Smoothness Constraints from Discrete Samples
Neural Computation, 1990
A comparison between CMAC neural network control and two traditional adaptive control systems
IEEE Control Systems Magazine, 1990
Real-time application of neural networks for sensor-based control of robots with vision
IEEE Transactions on Systems, Man, and Cybernetics, 1989
Parallel structures in human and computer memory
AIP Conference Proceedings, 1986
A New Approach to Manipulator Control: The Cerebellar Model Articulation Controller (CMAC)
Journal of Dynamic Systems, Measurement, and Control, 1975