Spectral-Distance-Based Detection of EMG Activity From Capacitive Measurements

Electromyographic (EMG) activity is often used as a diagnostic tool in hospitals. In addition, in the last decade, EMG signals are used for the control of assistive devices such as active exoskeletons and prostheses. Conventional surface electrodes used for the acquisition of EMG signals can cause irritation, discomfort, skin allergy, and inflammation. To cope with these issues capacitive contactless EMG electrodes have been developed. The method proposed in this paper is based on the use of short time Fourier transform and spectral distance computation to detect muscle activation from signals recorded using capacitive sensors. The method has been experimentally tested on five young healthy subjects. To assess the reliability of our method, we quantified the latency between the output of the detection algorithm and results of a gold standard protocol created by an expert bioengineer. The results show reliable off-line EMG detection. In the future, we plan to test whether the method could be used for online EMG detection.