Navigation of autonomous robots with an intelligent oscillator controller

Abstract

This paper describes a nonlinear oscillator scheme to control wheeled robots. The method is based on studies of control paradigms found in the natural world, With this method, the oscillator state is translated into motor drive signals. A sensor fusion network combines information about the state of the robot and environment with knowledge of the robot's goals in order to produce an oscillator driving force. The performance of this control scheme is examined in two different mazes, where the goal is to quickly explore all areas of the maze. It is shown that the controller can perform effectively in these environments.