Source position estimation via subspace based joint sparse recovery

This paper addresses the multi-source localization problem by utilizing a novel subspace based joint sparse recovery approach. We firstly introduce a multiple measurement vector (MMV) positioning framework in the context of joint sparse recovery. To sufficiently degrade the effect of measurement noise at a lower cost of deployed anchors, we then optimally hybridize the compressive sensing and array signal processing so that a part of location supports are first identified by CS scheme and the remaining supports are determined by a generalized subspace criterion. Meanwhile, we design a post refined procedure to further improve the estimate accuracy by considering the grid assumption. Finally, a comprehensive set of simulations had been conducted to demonstrate the superiority of our method over the traditional MMV algorithms.