Dynamic compressed sensing based track-before-detect algorithm for dim target tracking

This paper presents a novel dynamic compressed sensing based track-before-detect (DCS-TBD) algorithm. The proposed DCS-TBD algorithm reconstructs the radar scenario at each scan, and extracts the states of potential targets from the reconstructed radar scenario. The proposed algorithm distinguishes the true targets from clutter via a Kalman filter based potential targets management scheme. Following this, the proposed algorithm sends the predicted states of the true targets to the dynamic compressed sensing framework as prior information of the next scan. Finally, it is demonstrated that the proposed DCS-TBD algorithm is capable of track dim targets effectively.