Neural Oscillator Based Motion Control System for Snake-like Robot

Abstract

Robustness in the motion and structure is an important issue for mobile robots in order to realize specially, practical robots in open fields. A robust system which is strong enough to disturbance and breakdowns should be introduced into the robot system. As a control algorithm for a robust control system, a bio-inspired system imitating a nervous network system is considered as a candidate for the issue. In spiral cord of animal, a rhythm generator mechanism called the Central Pattern Generator (CPG) exists. The CPG consists of many neural oscillators with inhibitory connection, and neural oscillators influence each other and produce make rhythmical patterns. In this paper, a neural oscillator based motion control system is developed and applied to the motion control of a snake-like robot. The robot can move forwards, backwards by changing weights and change the heading direction by adjusting command parameters from the upper layer.