CPG-based locomotion control of a snake-like robot for obstacle avoidance

Abstract

This paper presents a biomimetic approach based on central pattern generator (CPG), to control turning motion of a snake-like robot. One of the interesting features of a biological snake is its ability to avoid obstacles or a barrier by turning its whole body from its trajectory. This special obstacle avoidance motion is different from other types of animal, and thus, it is worth to be analyzed and realized into a snake-like robot. The paper first briefly explains: 1) the phase oscillator model which represents the CPG model and 2) the CPG network. Next, we address several issues related to the existing/typical turning control of a snake-like robot. We then propose the phase transition method utilizing the phase difference control parameter to realize the turning motion of a snake-like robot. We also introduce a new parameter to control the turning of the robot, where it provides a way to incorporate sensory feedback into the CPG model. Simulation results show that the proposed turning method can be used efficiently as an obstacle avoidance method for a snake-like robot.