Analysis of Human COM Motions toward Improving Guidance of a Dance Teaching Robot

Improving synchronized motions in physical human-robot interactions

抄録

Real time estimation of human body motion is usually needed to understand and react in physical human-robot interactions (pHRI). In this paper, we present an online estimation of the human's COM motion in a pHRI with a robot teaching dance. The robot estimates the synchronization error and uses it, toward improving the performance in a dance training scenario lead by the robot. The method developed to obtain all parameters of the human body is based on data from laser range finders and a model of the human body in contact with the robot from motion capture data. This estimation of the COM is later used for understanding the student's current ability level, in this way, improving the robot response during the guidance in the learning experience.