Mobile Robot Navigation based on Human Walking Trajectory in Intelligent Space

Abstract

This paper presents a mobile robot navigation method based on human walking trajectory in intelligent space. Unlike the traditional navigation methods, the proposed method enables the robot to walk a meaningful path, which reflects human activity patterns in daily life. Human walking trajectories are extracted using multiple distributed vision cameras in intelligent space. To group trajectories with the same direction, trajectory similarity measurement is implemented. Key points on human walking trajectories and typical transfer points are utilized to build a topological map. To navigate a mobile robot to walk along the trajectory of human walking, global path planning based on the topological map and local path planning based on gird map are carried out with the spared information in intelligent space. Experimental results illustrate the performance of the proposed navigation method.