Comparison of Two Compressive Sensing Algorithms for Automotive Radar

Abstract

We analyze the possibility of using compressive sensing algorithms for the rapid chirps waveform in automotive radar sensor applications. Two algorithms are considered: Orthogonal Matching Pursuit (OMP) and l1-magic. We compare the two methods in a scenario using nine targets, with and without noise. The number of non-uniformly placed samples is four times less than the number of uniform samples, with target detection possible even in the presence of noise. It is shown that OMP outperforms l1-magic, with spectra not affected by the Compressive Sensing reconstruction. The results are comparable with the traditional uniform sampling results.