Locomotion strategy selection for a legged wheeled hybrid quadruped using depth images

Abstract

Three fundamental locomotion configurations recognized commonly are legged, wheeled, and articulated mechanisms using which a mobile robot can navigate terrains. Hybrid configurations enable execution of different locomotion types separately and in combinations. Such advantage usually implies complexity and necessity in a robust supervisory controller capable of terrain recognition and locomotion strategy selection. We developed the Nazarbayev University (NU) Hybrid Quadruped (Fig. 1) - mobile robot with four legs and wheels. Project's major novelty is the implementation of the supervisory controller which selects a locomotion mode associated with particular terrain types based on its terrain recognizer input data.