Geometric feature-based footholds identification for an insect-like walker

Legged systems theoretically offer the potential to better traverse rugged terrains than traditional wheeled and tracked designs due to no need for paths of continuous contact with the ground, and therefore have been given great concern. In this paper, a geometric feature-based footholds identification system for a small six-legged robot HITCR-II was presented. For this algorithm, a novel local terrain feature extraction method was introduced, taking account of the shape and size of the robot foot. A ranking function used to quantify the risk of a candidate location for foothold by the robot was designed and trained using support vector machine (SVM). Experimental results showed that the proposed footholds identification system was able to successfully rank the usefulness of the ground points.