Automatic Adjustment of Electromyography-Based FES Control

Cologni A.L., M.G. Madaschi, F. Previdi, C. Werner, T. Schauer

In this work, two EMG-based FES control approaches are investigated with the aim of allowing stroke patients to recover motor function of the upper limbs, especially the wrist. Often, patients have weak residual muscle activity but are not able to achieve high wrist extension angles. This residual volitional muscle activity can be continuously measured from the stimulation electrodes by using a customised EMG amplifier. The aim of both developed control approaches is to find the optimal FES support that uses the minimal required stimulation intensity to support the movement and  that yields an almost linear relation between the measured volitional EMG activity (effort) and the resulting wrist-joint angle. The first approach is based on iteratively learning a static EMG-stimulation intensity relation from repeated EMG-reference tracking tasks, while the second approach maps the detected volitional EMG activity to a reference angle that is fed to a feedback controller. Both cases require the measurement of the wrist-joint angle, e.g. by an inertial motion unit. For both techniques, case studies with healthy subjects and stroke patients have been conducted. Advantages and disadvantages of both approaches are discussed in this contribution