Rapid Reinforcement Learning for Reactive Control Policy Design in Autonomous Robots

Abstract

A key property of Artificial Life systems is their ability to autonomously learn from experience while behaving within a dynamic environment. By autonomous learning, we mean that a system is capable of acquiring control programs with little information provided by some external teacher. To this end, we have developed a neural-based reinforcement learning architecture for the design of reactive control policies for an autonomous robot. Reinforcement learning techniques allow a programmer to specify the control program at the level of the desired behavior of the robot, rather than at the level of the program that generates that behavior. In this chapter, we address the issue of state representation which can greatly affect the system's ability to learn quickly and to apply what has already been learned to novel situations. Finally, we demonstrate the architecture as applied towards a real robot that is learning to move safely about its environment. Introduction Traditional methods of co...