Energetic Passivity Decoding of Human Hip Joint for Physical Human-Robot Interaction

Abstract

The capacity of the biomechanics of human limbs to absorb energy during physical human-robot interaction (pHRI) can play an imperative role in controlling the performance of human-centered robotics systems. Using the concept of “excess of passivity,” we have recently designed passivity signature maps for elbow and wrist joints. We have also shown that this knowledge can be exploited and extrapolated during the interaction with a robotic system by transparency-maximized algorithms. A major application is in robotic rehabilitation systems and assistive technologies. Here, for the first time, the nonlinear energy capacitance of the hip joint and the affecting factors are decoded. This can be critical for maximizing the performance of wearable exoskeletons. Knowledge regarding energy absorption behavior can significantly help to reduce the conservatism of control algorithms. In this work, the energetic behavior is studied for three different hip angles, while perturbations were provided at three different interaction speeds. The results show that the increase in agonist and antagonist muscle contractions can consistently expand the margins of the passivity map. Additionally, by separating the effects of agonist and antagonist contractions, it was identified that the passivity margins have a correlation with the subject's posture during interaction with the robot and the correlation depends on the type of muscle contraction. A preliminary design of a stabilizer is also formulated that takes into account variable passivity behavior of the joint, in the energy domain, to enhance the performance while guaranteeing pHRI stability.