Underdetermined blind source separation for LFM radar signal based on compressive sensing

Blind source separation (BSS) problem for ultra wide band (UWB) linear frequency modulated (LFM) signals under underdetermined case (the case of less observed signals than sources) is discussed and the framework based on compressive sensing (CS) to solve this problem is presented. In the first step, a mixing matrix recovery algorithm based on the normal vector of hyperplanes is given. In the second step, for LFM signals, sparse dictionary is designed as part of the compressive sensing process. To reconstruct the sources, the simple orthogonal matching pursuit (OMP) is chosen with less data storage and lower computational complexity. Finally, simulations are taken on testing the proposed framework on LFM signals, and the results are provided to verify the feasibility and efficiency of the novel method.