Research on an SVM-Based Adaptive Sliding Mode Composite Control Strategy for Lower Limb Rehabilitation Under Non-Periodic Gait Conditions

Abstract

The old way of controlling lower-limb prostheses is a fixed-period gait model; it cannot adapt to various walking conditions in reality and is also relatively unstable at the beginning of rehabilitation. Given the problems mentioned above, an SVM-based adaptive sliding mode hybrid control method will be put forward in this paper. Therefore, the above method will be employed to correct the non-linear error by means of SVM, and then adaptive sliding mode control will be used to improve stability and tracking performance. Simulations in MATLAB/Simulink of the S5 combined disturbance scenario, which includes harmonic torque perturbation, ground impact and sensing noise, have shown that the proposed scheme can reduce the RMSE values of the hip, knee and ankle from 1.371°, 2.314° and 0.970° to 0.355°, 0.789° and 0.290°, respectively; that is to say, there is an improvement of 74.1%, 65.9% and 70.1% over the proportional-derivative (PD) reference controller. The new way will also reduce the peak error of impact and improve the stability of non-periodic walking by quickly changing the path. Control Architecture of an Intelligent Lower-Limb Rehabilitation Exoskeleton Device is Presented in This Paper.