Decentralized Position/torque Control of Modular Robot Manipulators via Interaction Torque Estimation-based Human Motion Intention Identification

For the application background of physical human robot interaction (pHRI), a novel decentralized position/torque control scheme of modular robot manipulators (MRMs) is developed based on the human motion intention identification in this investigation. Different from traditional control schemes which are oriented to pHRI tasks depending on the biological signal or the multisensory, the developed decentralized position/torque control is realized by utilizing only position measurements of each joint module in this paper. A novel extend state observer (ESO)-based interaction torque estimation method is proposed to identify the human motion intention that can provide motion tracking information for robots in pHRI tasks by using only joint local dynamic information. In addition, the interaction torque estimation is utilized to design the decentralized position/torque control scheme which can implement high-performance of interaction torque tracking and position tracking. The trajectory tracking error and the interaction torque tracking error of the closed-loop MRM system are uniformly ultimately bounded (UUB) which is proved by Lyapunov theory. Finally, pHRI experiment is utilized to verify the effectiveness and advancement of the proposed method.