Walking Trajectory Planning on Stairs Using Virtual Slope for Biped Robots

Abstract

In this paper, a “virtual slope method” for walking trajectory planning on stairs for biped robots is proposed. In conventional methods for walking on stairs, there are two problems about the zero-moment point (ZMP). One is a ZMP equation problem, and the other is a ZMP definition problem in a double-support phase. First, a ZMP equation on stairs is different from that on flat ground. Therefore, the same trajectory generation as flat ground cannot be implemented. This problem is defined as a “ZMP equation problem.” Second, the ZMP cannot be defined in the double-support phase on stairs because contact points of the feet do not constitute a plane. The ZMP can be defined only on the plane. This problem is defined as a “ZMP definition problem.” These two problems are solved concurrently by the virtual slope method. It is the method that regards the stairs as a virtual slope. In walking trajectory planning on a slope of the constant gradient, the two problems about the ZMP do not exist. Additionally, a trajectory planning procedure based on the virtual slope method is explained. The validity of the proposed method is confirmed by some simulations and experiments.