Fast Recovery of Low-Rank and Joint-Sparse Signals in Wireless Body Area Networks

E-health monitoring signals collected from wireless body area networks (WBANs) usually have some highly correlated structures in a certain transform domain (e.g., discrete cosine transform (DCT)). We exploit these structures and propose a fast recovery algorithm for low-rank and joint-sparse (L&S) structured WBAN signal in the framework of compressed sensing (CS). By using a simultaneously L&S signal model, we employ the number of the bigger singular values and Bayesian learning which incorporates an L&S-inducing prior over the signal and the appropriate hyperpriors over all hyperparameters to recover the signal. Experiments show that the proposed algorithm has a superior performance to state-of-the-art algorithms.