Soft Iterative Method with Adaptive Thresholding for Reconstruction of Radar Scenes

Abstract

A proper use of the advances of compressed sensing (CS) theory in radar systems may lead to data rate reduction, energy saving and measuring improvement. Using, e.g., randomly placed chirps and CS reconstruction algorithms, the velocity unambiguously measured by FMCW-Radars can be increased. In this paper, we propose to use the Soft Iterative Method with Adaptive Thresholding (soft-IMAT) for the reconstruction of the radar scene. Its performance is analyzed in different scenarios and compared with the original IMAT with respect to detection probability and computational complexity. A modified soft-IMAT is proposed in order to reduce the computational overhead caused by the soft-based decision.