In a wireless body sensor network (WBSN), the available energy and bandwidth are limited. Therefore, compressing the electromyogram (EMG) signal is of great importance since it is generally sensed at a relatively high frequency of the order of kHz. In this paper, we use the compressed sensing (CS) technique to compress and recover the EMG signal. The main advantage with CS is that its compression process requires less computational complexity. We propose a deterministic measurement matrix that greatly facilitates the implementation of the encoder device. The simulation and experiment results showed that the proposed approach can compress and recover the EMG signal without perceptible loss if the compression ratio was greater than or equal to 0.25, which saved up to 75 % of both the available bandwidth and power consumption of the transceiver. A comparison with the current stat-of-the-art of EMG compression shows that we obtained a better performance. Furthermore, the proposed encoder has the lowest computational complexity.
展开▼
