Backstepping Sliding Mode Control With Enhanced Error-Dependent Observer Switching Mechanism for Knee Exoskeleton

Exoskeletons are currently widely used for athletic enhancement and rehabilitation training. However, the force loading accuracy of wearable flexible knee exoskeletons is limited by system characteristics and the human-machine interaction. In this paper, an effective force loading control algorithm is proposed. First, a detailed human-machine interaction system model of the exoskeleton is established. Second, an enhanced error-dependent observer switching mechanism is proposed for the exoskeleton force loading system. The stability of a nonlinear extended state observer (NESO) for a third-order force loading system is rigorously proved using the Lyapunov theorem, thereby addressing the difficulty of analyzing the stability of the switched extended state observer (SESO). Based on the SESO, a backstepping sliding mode controller (BSSM) is designed to ensure that the dynamic performance of the force loading is improved with a small switching gain. Finally, comparative experiments verify the effectiveness of the proposed control method, which can effectively suppress parameter perturbations and external disturbances, thereby achieving high-precision tracking.