Compressed sensing of respiratory signals promoting joint-sparsity

Abstract

Telemonitoring is a potential solution for management of patients suffering from chronic respiratory diseases, such as chronic obstructive pulmonary disease (COPD), respiratory failure, and obstructive sleep apnea. However, the compression is a prime concern for designing telemonitoring systems via Wireless Body Area Networks (WBANs). In this regard, Compressed Sensing (CS) is a promising tool of compression. This paper proposes a mixed norm-based CS technique to compress/recover respiratory signals in WBAN systems. To enhance the recovery performance, the overall problem is framed in Multiple Measurement Vector (MMV) model exploiting the joint-sparsity. First, the raw respiratory data is compressed employing a sparse binary sensing matrix with a few nonzero entries at the sensor (transmitter) side. Then at the server (receiver) side, the original signal is recovered using the proposed algorithm. The experimental results using the Physiobank respiratory database shows promising achievement obtained by the proposed method in terms of CPU computation time as well as reconstruction quality.