Flexible assistance strategy of lower limb rehabilitation exoskeleton based on admittance model

Exercise training based on exoskeleton is an effective rehabilitation method for stroke patients. However, some rehabilitation exoskeletons still have poor wearable performance and obvious human-robot impedance during training, which easily cause secondary injuries to the patients. In this study, a variable admittance control strategy is proposed to improve the operator’s wearable comfort, which can adapt to different operators by regulating admittance parameters. The admittance controller has two feedback loops: the position inner-loop and the admittance outer-loop. Meanwhile, the exoskeleton Lagrange model is constructed with unknown friction disturbance. By using Lyapunov technique and backstepping with state observer, the joint position error of exoskeleton is convergence into a zero neighborhood. The effectiveness of the proposed admittance controller is verified by both simulation and experiment. The tracking error of hip and knee joint is less than 4 degrees while the human-robot interaction torque is constrained in a tolerable range of the operator.