A self-adaptive foot-drop corrector using functional electrical stimulation (FES) modulated by tibialis anterior electromyography (EMG) dataset

We developed a functional electrical stimulator for correcting the gait patterns of patients with foot-drop problem. The stimulating electrical pulses of the system are modulated to evoke contractions of the tibialis anterior muscle, by emulating the normal patterns. The modulation is adaptive, i.e. the system can predict the user's step frequency and the generated stimulation can match each step in real-time. In this study, step data from 11 young healthy volunteers were acquired, and five prediction algorithms were evaluated by the acquired data, including the average of Previous N steps (P-N), the Previous Nth step (P-Nth), General Regression Neural Network (GRNN), Autoregressive (AR) and Kalman filter (KF). The algorithm with the best efficiency-accuracy trade-off (P-N, when N=5) was implemented in the FES system. System evaluation results obtained from a post-stroke patient with foot-drop showed that the system of this study demonstrated better performance on gait pattern correction than the methods widely adopted in commercial products.