Sparse Detection Algorithms Based on Two-Dimensional Compressive Sensing for Sub-Nyquist Pulse Doppler Radar Systems

TL;DR: Four 2-D compressive sensing algorithms which are extended from the traditional 1-D CS algorithms (the ZAP, IHT, ISTA, and FISTA algorithms) are proposed which can achieve comparable detection performance with lower memory requirement and eliminate the interference of impulsive noise in the non-Gaussian impulse noise environment.

Abstract: Sub-Nyquist pulse Doppler radar system has received wide attention recently because it can make the sampling rate lower than the Nyquist sampling rate. However, there are still two problems that must be addressed for sub-Nyquist pulse Doppler radar. One is that the large memory is required for sparse signal recovery. Another is that the performance of the sparse signal recovery will be distorted in the non-Gaussian impulse noise environment. For the first problem, this paper proposes four 2-D compressive sensing (CS) algorithms which are extended from the traditional 1-D CS algorithms (the ZAP, IHT, ISTA, and FISTA algorithms). The proposed 2-D-CS algorithms recover the signal in the delay-Doppler domain, which is a matrix domain. For the second problem, robust 2-D-CS algorithms are proposed for the non-Gaussian impulse noise environment. The proposed 2-D-CS algorithms can achieve comparable detection performance with lower memory requirement. The proposed robust 2-D-CS algorithms can eliminate the interference of impulsive noise in the non-Gaussian impulse noise environment. Simulation results are given to verify the effectiveness of the proposed algorithms.