Target Human Detection Based on Matching of Walking Motion Signals Between Smartphone and Robot for Human Following

Abstract

This paper introduces a target human following mobile robot system. Especially, the robot follows a smartphone user as the target. The proposed system includes global following based on Wi-Fi-based waypoint maps and local following based on walking state matching. This paper focuses on target human detection for local following based on cross-correlation coefficients of walking motion signals including vertical acceleration data measured by target user's smartphone and horizontal velocities of legs measured by an LRF installed in the robot. Also, this paper provides results of detection experiments. The results show that higher cross-correlation coefficients can be obtained for the target human than the other pedestrians. Next, a method for calculating the cross-correlation coefficients online is introduced, and the results of the experiment are shown. These results show that the following target can be identified in online calculation of cross-correlation coefficients. That contributes for development of actual human following system.