Approach for Adaptive Assistance Control of Lower Limb Rehabilitation Robot According to Human Need

Abstract

The control of lower limb rehabilitation robot mostly is lacking in human active participation during gait rehabilitation, which can not provide suitable assistance based on the human impairment level. A novel adaptive assistance control strategy according to human need is presented for lower limb rehabilitation robot. The rehabilitation degree of human subjects is learnt in real time following the tracking errors between the man-machine movement trajectory and the ideal rehabilitation trajectory, so that the rehabilitation robot can adaptively provide the assist-as-needed. The man-machine dynamic model of the lower limb rehabilitation robot system is established and the assist-as-needed controller of lower limb rehabilitation robot is designed. The simulation for the proposed algorithm is conducted and the experiment with healthy subjects is implemented on single mechanical leg prototype. The results show that the assist-as-needed controller can adaptively attenuate the assist torque of rehabilitation robot according to human rehabilitation level as long as the forgetting term remains non-zero to serve the purpose of assistance according to human need.