A New Method for sEMG Envelope Detection from Reduced Measurements

The amount of data to be transmitted from each smart device to the cloud server increases with the number of sensors, so compressing the acquired biosignals before transmission is relevant to increase the efficiency in Internet of Things networks. This principle applies to surface Electromyography (sEMG) signals for gait analysis as well. The paper proposes a new method based on Compressed Sensing (CS) for sEMG processing from reduced measurements. A deterministic matrix is chosen to model the compression phase. Instead, a matrix built with a Daubechies wavelet kernel is considered for the reconstruction phase. The CS reconstruction is then applied to the detection of a significant feature of sEMG signals, that is the linear envelope. Thus, the CS-based method for envelope detection is analyzed on sEMG signals corresponding to strides measured by 10 healthy subjects. The proposed method proves to be reliable for envelope detection. In fact, by comparing peak amplitudes and time positions of envelopes corresponding to reconstructed and original signals, the proposed CS-based method shows an irrelevant loss of information