Moving targets extraction in multichannel wide-area surveillance system by exploiting sparse phase matrix

A novel algorithm is proposed to extract moving targets using the phase information of the radar echoes in multichannel wide-area surveillance radar system. The algorithm is based on the recently proposed relaxed version of principal component pursuit (PCP) method, which is for the first time introduced into radar signal processing community. In the algorithm, after pre-processing and channel combination, the combined phase matrix of radar echoes can be expressed as the superposition of three matrices: low-rank phase matrix of ground clutter, sparse phase matrix of moving targets and entry-wise phase matrix of the noise component. This expression fits well with recently proposed the relaxed version of PCP method. Therefore accelerated proximal gradient algorithm can be applied to stably extract sparse phase matrix of moving targets. Both simulation and real data processing results verify the effectiveness of this proposed new algorithm.