A walking motion imitation framework of a humanoid robot by human walking recognition from IMU motion data

Abstract

Teleoperation of humanoid robots, which is one of good methods of humanoid utilization, has difficulties in locomotion with self-balancing. In this paper, we propose a framework of walking imitation between human and a humanoid robot. Using inertial measurement unit (IMU), human's walking motions are recognized, and used as humanoid robot control inputs for on-line walking imitation. Appropriate IMU feature data for walking recognition are investigated, and walking recognition method based on supporting foot detection and stance estimation is suggested. A humanoid robot is modeled as an inverted pendulum for on-line walking control. For validation, experiments were conducted using Mahru-R, a humanoid robot developed in KIST.