Development of collision-free-area detection algorithm for mobile robot

Abstract

Detection of collision-free area for mobile robots is one of the important functions of safe autonomous navigation. This paper describes the development of a collision-free area detection algorithm for an autonomous mobile robot using a depth camera. The depth camera can simultaneously detect three-dimensional shapes of obstacles around the mobile robot, similar to a conventional vision sensor. However, the depth camera detects not only obstacles but also ground surface slopes, even if the mobile robot can safely move. To detect a collision-free area, regardless of the ground surface slopes, we develop a new collision-free-area detection algorithm for mobile robots. The validity of the proposed algorithm is confirmed by simulations and actual outdoor experiments.