Signal selective 2-D direction finding for non-Gaussian cyclostationary sources

In this paper, we present a new 2-D direction-of-arrival (DOA) estimation algorithm by exploiting the merits of cumulant and cyclostationarity. The proposed approach constructs a new matrix called virtual DOA cyclic matrix (VDOA Cyclic Matrix) in cyclic cumulant domain to estimate 2-D DOA with two virtual parallel linear arrays. The theoretical analysis shows that the 2-D DOA of the sources can be found through the eigen-decomposition of VDOA Cyclic Matrix requiring no search procedure to correctly pair 2-D angles. The algorithm is featured with array aperture extension and signal selectivity, so that the direction finding performance is enhanced