Pediatric gait training using a lower-limb exoskeleton with adaptive finite-time sliding mode control scheme: An experimental study

Abstract

This study addresses challenges in real-time gait regulation for a pediatric exoskeleton by introducing an adaptive finite-time sliding mode control. The approach utilizes a non-singular fast terminal sliding mode (NFTSM) surface with adaptive laws to guide the system states toward equilibrium within a finite time. Lyapunov’s theory validates the swift convergence of tracking error. The proposed adaptive-NFTSM (ANFTSM) control is applied on an existing lower-limb exoskeleton and tested for a Cerebral Palsy-affected pediatric subject (12 years, 35 kg, 131 cm) over 7 days of gait training. On the seventh day, ANFSTM control strategy outperforms the contrast adaptive terminal sliding mode (ATSM) strategy by 21.57% and 16.36% in tracking knee and ankle joint trajectories. ANFSTM control achieves convergent sliding manifolds $(|s{|}_{\mathrm{hip}}<0.01,|s{|}_{\mathrm{knee}}<0.05,|s{|}_{\mathrm{ankle}}<0.02)$ within 25%, 23%, and 21% of the respective gait cycles. Moreover, the proposed ANFTSM control demonstrates negligible chattering in sliding surfaces over 7 days compared to the adaptive terminal sliding mode (ATSM) control. The precise estimation of unknown coefficients in the ANFTSM control ensures robustness against uncertainties and external interferences. On the seventh day, ANFTSM control excels over ATSM control with a remarkable reduction of more than 18% in voltage consumption. Finally, statistical measures over the entire gait training period emphasize the potential of the ANFTSM control scheme in ameliorating abnormal gait patterns in pediatric subjects.