Estimating 2-D angles of arrival via two parallel linear arrays

A novel approach for high resolution estimates of two-dimensional (2-D) angles of arrival of multiple narrowband sources for two parallel linear arrays is presented. By fully using the properties of the autocovariance matrices, the authors construct a direction-of-arrival (DOA) matrix. The theoretical analysis shows that the eigenvalues and corresponding eigenvectors of the DOA matrix are related to the 2-D angles of arrival of sources. Thus, they can be used to estimate 2-D angles of arrival separately, and the large computations required for 2-D searches can be avoided. Simulation results are given, showing that the method yields excellent estimates with low variance