An approach to direction of arrival estimation of multiple coherently distributed sources by searching maximum eigenvalue

To circumvent the problem of performance degradation of direction of arrival (DOA) estimation due to angular spread, it was assumed impracticably that the parameterized shape of the angular distribution is known. Moreover, the disadvantage of most algorithms is the computational complexity as a multi-dimensional numerical search is necessary. By using a simple 1-D maximum eigenvalue search rather than resort to multidimensional parametric search, we propose a blind algorithm for DOA estimation of multiple coherently distributed sources, without prior knowledge about the function form of the angular distribution. We develop the algorithm in uniform linear array (ULA) condition, although our approach is not limited to ULA.