Trajectory tracking using artificial neural network for stable human-like gait with upper body motion

Abstract

This paper presents a trajectory generation algorithm for robots which can walk like human with movable foot and active toe. The proposed algorithm allows smooth transition between walking phases namely, single and double support phases. A neural network approach is used for solving inverse kinematics so that the biped robot follows the ankle and hip trajectories to walk. Zero moment point (ZMP) stability is ensured by taking into account the upper body movements along with the planned motion trajectories. Here, we analyze the effect of lateral upper body motion on ZMP stability. Different types of trajectories for upper body are generated, and the one which ensured the most stable locomotion is identified.