Adaptive backstepping human-cooperative control of a pediatric gait exoskeleton system with high- and low-level admittance

Post-neurological disorder, passive-assist training disregards human involvement during robot-aided lower-limb rehabilitation. This work presents a novel human-cooperative framework based on the admittance and trajectory control scheme to administer the subject-exoskeleton interaction for pediatric gait rehabilitation. Initially, the mechanical design and dynamic analysis of an existing exoskeleton system are briefly attended. Thereafter, an admittance model is designed in the outer loop, which recasts the desired human trajectory into a reference trajectory. As an inner-loop control scheme, a robust adaptive backstepping controller is designed to trace the reference gait trajectory under parametric uncertainties and external disturbances. Lyapunov stability analysis is solved to guarantee the uniform boundedness of the control signals. Moreover, the well-known problem of "explosion of complexity" and "overparameterization" is avoided through the design of the robust adaptive backstepping control scheme. The performance of the proposed robust adaptive backstepping-based human-cooperative control is studied under the low- and high-level admittance model. Finally, the effectiveness of the proposed control is validated with a variable structure adaptive robust-based human-cooperative control. The co-simulation results show that the proposed control with low-level admittance allows the subject to participate in the training process more frankly. The proposed robust adaptive backstepping-based human-cooperative control tracks the reference gait more promisingly than the variable structure adaptive robust-based human-cooperative control for both admittance levels.