Motion Retargeting for Humanoid Robots Based on Simultaneous Morphing Parameter Identification and Motion Optimization

This paper presents a novel method for retargeting human motions onto a humanoid robot. The method solves the following three simultaneous problems: the geometric parameter identification that morphs the human model to the robot model, motion planning for a robot, and the inverse kinematics of the human motion-capture data. Simultaneous solutions can imitate the original motion more accurately than conventional approaches, which solve the problems sequentially. The proposed method can reconstruct the human motion within the physical constraints imposed by robot dynamics. A reconstruction step enables quantitative analysis of the retargeting results through direct comparison with the original human motion. The method can also provide the precise morphing function as well as subject-specific models, which can handle the different body dimensions of human subjects. This new framework is suitable for applications that require an accurate generation of human-like motions with quantitative evaluation criteria, such as humanoid robots that evaluate assistive devices. Experimental tests of the proposed method were performed with humanoid robot HRP-4.