Terrain Classification, Navigation, and Gait Selection in a Leg-Wheel Transformable Robot by Using Environmental RGBD Information

Abstract

This paper presents a behavior decision strategy for the leg-wheel transformable robot, TurboQuad-V. By using real-time RGBD information combined with the gait selection policy concluded from the single-leg-wheel dynamic simulation, the robot can perform autonomous gait switching and basic obstacle avoidance on multiple terrains.