Command filtered backstepping control with nonlinear disturbance observer for multi-posture lower limb rehabilitation robot

This paper proposes a novel command-filtered backstepping control method integrated with nonlinear disturbance observers (NDO-CFBS) to improve early-stage rehabilitation training for patients with lower-limb dysfunction across multiple postures, including sitting, lying, and standing. A backstepping control strategy is first developed for the human–robot system, considering the characteristics of passive rehabilitation training. Nonlinear disturbance observers (NDOs) are designed to compensate for model uncertainties and external disturbances. A command filter is then introduced to efficiently compute the derivatives of the virtual control variables within the backstepping framework, leading to the formulation of the NDO-CFBS method. Finally, simulation results of trajectory tracking performance are presented for the multi-posture lower limb rehabilitation robot, demonstrating the effectiveness of the proposed control strategy.