A geometric approach to gait generation for eel-like locomotion

Abstract

We investigate issues of control and motion planning for a biomimetic robotic system. Previous work has shown that the motion problem can be decoupled into trajectory generation and steering. We investigate basic issues of momentum generation for a class of dynamic mobile robots, focusing on eel-like swimming robots. We use control laws based on a series of gaits motivated by the biological literature on coupled oscillators and central pattern generators. A primary characteristic of this class of robots is that drift plays a significant role in the generation of motion. We develop theoretical justification for a forward gait that has been observed in nature and for a turning gait, used in our control laws, that has not yet been studied in the biological literature. We also explore theoretical predictions of novel gaits for turning and sideways swimming.