Biologically inspired autonomous robot that learns approach-avoidance behaviors

Abstract

This paper describes the design of a robotic entity that is capable of independent action in open, unpredictable environments. In order to accomplish this goal we search for inspiration in biological systems. Yet we do not attempt to blindly replicate all specific details (of implementation) since many of them may be accidental. But we attempt to extract general principles of organization that seem to have proved essential for the success of biological autonomous systems. The result is a robust, flexible and adaptive control system for the autonomous robot.