An Adaptable Oscillator-Based Controller for Autonomous Robots

  • Authors:

  • Gary T. Anderson;Youlong Yang;Gang Cheng

  • Affiliations:

  • Department of Applied Science, University of Arkansas at Little Rock, Little Rock, USA 72204-1099;Department of Applied Science, University of Arkansas at Little Rock, Little Rock, USA 72204-1099;Department of Applied Science, University of Arkansas at Little Rock, Little Rock, USA 72204-1099

  • Venue:
  • Journal of Intelligent and Robotic Systems
  • Year:
  • 2009

Quantified Score

Hi-index 0.00

Visualization

Abstract

This paper proposes a unique oscillator-based robot controller with learning abilities to effectively guide a team of robots operating in uncertain environments. To verify this, we designed four separate controllers and compared their performance in a series of tests in several different environments. The experiments used a team of three robots to explore arenas with variable lighting and different obstacle patterns, with a goal of having the team as a whole absorb as much light as possible. The four controllers were: a reactive controller, an oscillator with fixed parameters, an oscillator whose parameters changed based on the pattern of sensor information received, and an oscillator-based controller that used reinforcement learning. Experiments confirmed that the proposed method outperforms the others in all environments tested.