Variable Impedance Decentralized Control of Reconfigurable Robot Manipulator for Human Robot Collaboration

A variable impedance decentralized control method of reconfigurable robot manipulator (RRM) with human robot collaboration (HRC) using Gaussian process-based motion intention estimation is proposed in this paper. The RRM system dynamic model based on joint torque feedback technology and the human arm dynamic model based on mechanical impedance are established. The human motion intention is identified by Gaussian process regression (GPR) method. A variable impedance decentralized control strategy is designed to adjust the damping coefficient in real time according to the HRC force to make RRM conform to human motion intention. The decentralized integral terminal sliding mode control is adopted to further improve the position tracking accuracy. Finally, based on Lyapunov theory proves that the RRM tracking error for HRC task system is ultimately uniformly bounded.