A comparison of CS reconstruction algorithms for multicomponent nonlinear phase signals

This paper analyzes the performance of different compressive sensing algorithms applied to signals with polynomial and cosine modulated phases, that usually appear in radar communications. In order to provide sparsity in the Fourier transform domain, the signal components are firstly demodulated by a direct parameter search method. In this way, the signals are sparsified in the DFT domain. The performance of reconstruction algorithms are compared using small number of measurements. The results are illustrated on examples.