Pneumatic Artificial Muscle Actuated Robot for Lower Limb Rehabilitation Triggered by Eelectromyography Signals Using Discrete Wavelet Transformation and Support Vector Machines

Abstract

In this study, a real-time electromyography (EMG)-triggered controller for a pneumatic artificial muscle (PAM) actuated lower limb rehabilitation robot is proposed. To make the rehabilitation task controllable by the patient's movement intention, a novel trigger controller is designed according to the EMG signals of the patient's muscle. For predicting his or her movement intention in advance, the EMG signals of the patient's muscle are captured and identified to realize the proposed EMG-triggered control. To guarantee the safety and performance of the proposed system, a patient's movement intention must be identified accurately by EMG feature extraction. First, the discrete wavelet transformation (DWT) technique is used to acquire the feature vectors of the EMG signals. The properties of the different feature spaces are taken into consideration, and the optimal multicomponents of features are chosen according to the experimental results. Second, support vector machines (SVMs) are studied to improve the classification performance. Finally, the MyRIO controller is used to implement a closed-loop control system for the rehabilitation robot with the movement-intention trigger control.