Learning momentum: online performance enhancement for reactive systems

Abstract

The authors describe a reactive robotic control system which incorporates aspects of machine learning to improve the system's ability to navigate successfully in unfamiliar environments. This system overcomes limitations of completely reactive systems by exercising online performance enhancement without the need for high-level planning. The goal of the learning system is to give the autonomous robot the ability to adjust the scheme control parameters in an unstructured dynamic environment. The results of a successful implementation that learns to navigate out of a box canyon are presented. This system never resorts to a high-level planner, but instead learns continuously by adjusting gains based on the progress made so far. The system is successful because it is able to improve its performance in reaching a goal in a previously unfamiliar and dynamic world.

Keywords

This publication has 8 references indexed in Scilit:

Learning reliable manipulation strategies without initial physical models
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2002
On Becoming Decreasingly Reactive: Learning to Deliberate Minimally
Published by Elsevier ,1991
Integrating behavioral, perceptual, and world knowledge in reactive navigation
Robotics and Autonomous Systems, 1990
Reducing Real-world Failures of Approximate Explanation-based Rules
Published by Elsevier ,1990
The impact of cybernetics on the design of a mobile robot system: a case study
IEEE Transactions on Systems, Man, and Cybernetics, 1990
Motor Schema — Based Mobile Robot Navigation
The International Journal of Robotics Research, 1989
A robust layered control system for a mobile robot
IEEE Journal on Robotics and Automation, 1986
An architecture for reflexive autonomous vehicle control
Published by Institute of Electrical and Electronics Engineers (IEEE) ,1986