New stability index for humanoid robot walking on rough terrain

Humanoid robots with two arms and legs are expected to act as substitutes for humans, and hence, they are required to move as humans do and accomplish the assigned tasks. Numerous studies have been conducted to stabilize the motion of humanoids. Two methods, Virtual Horizontal Plane (VHP) and Generalized Zero Moment Point (GZMP) are proposed. In both these methods, we consider a stability region; the robot remains stable as long as the defined ZMP is within this region. However, on uneven terrain the robot may not fall down even if ZMP lies outside the defined stability region. To address the issue, we focused on the internal force between a robot and its environment. In this paper, we propose a novel stability condition based on the normal component of the internal force and not the friction force. The validity of the proposed method is confirmed by a simulation and an experiment.