Extended State Observer-Based Nonlinear Terminal Sliding Mode Control With Feedforward Compensation for Lower Extremity Exoskeleton

This paper presents and experimentally demonstrates an extended state observer (ESO) -based nonlinear terminal sliding mode control strategy with feedforward compensation (ESO-F-NTSMC) for lower extremity exoskeleton. Since the lower extremity exoskeleton (LEE) is a coupled human-exoskeleton coordination system, the internal or external disturbances and uncertainties affect its performance. A nonlinear terminal sliding mode control with feedforward compensation (F-NTSMC) is proposed to drive the lower extremity to shadow the target human gait trajectory. An ESO is employed to estimate the total disturbances including these caused by the chattering phenomenon in F-NTSMC. ESO-F-NTSMC can assure that the human gait trajectory tracking can converge to a bounded region smoothly and robustly. The phase identification-based human gait generation approach is also presented. The derivation process of the ESO-F-NTSMC is shown the and the Lyapunov-based stability analysis is conducted. To illustrate the proposed method's effectiveness, experiments are performed on three human subjects walking on the floor at a natural speed. The results demonstrate that the exoskeleton can actively collaborate with the user under the proposed method.