Design and Test of Admittance Control with Inner Adaptive Robust Position Control for a Lower Limb Rehabilitation Robot

Although admittance control has been used in rehabilitation robots in many studies as it can realize compliant human-robot interaction, the inner proportional-integral-derivative (PID) controller of conventional admittance schemes is simple and not robust enough. This study presents an admittance control scheme with inner adaptive robust position control (ARC) for a hip-knee-ankle rehabilitation robot. The ARC is capable of eliminating uncertainties and external disturbances. A healthy male subject was required to perform three experiments, including passive exercise (PE) using a position controller, patient-cooperative exercise using an admittance controller (ACE) with fixed and different parameters. The PE experiment results show that the average normalized root mean square deviation (NRMSD) of trajectory tracking of each joint using the ARC is nearly 60% less than that using the PID controller. And the ACE experiment results show that the average NRMSD using the ARC is roughly 45% less than that using the PID controller while the interaction torque using the two controllers are comparable. It demonstrates that the robot becomes not only compliant but also robust based on the proposed control scheme. Moreover, the last experiment results indicate that admittance control with smaller admittance parameters allows the robot to be more compliant.