DOA estimation of cyclostationary signals in the presence of coherent sources

In multipath environments, cyclic MUSIC is failed to detect all incident sources. In order to de-correlate the coherent signals, spatial smoothing (SS) and forward/backward SS techniques have been investigated to resolve this situation, however, all these algorithms will reduce the number of detectable signals and cause the deficiency of array aperture. In this paper, two new approaches called virtual moving cyclic autocorrelation method (VM-CAC) and virtual moving cyclic crosscorrelation method (VM-CCC) are proposed for a uniform linear array (ULA). The theoretical analysis and simulation results indicate significantly better performance for the new methods in real-time implementation.