A DOA Estimation Algorithm Based on Variance Fusion with Double Uniform Linear Array

A fast high precise DOA finding algorithm based on double uniform linear arrays has been proposed in the paper. Based on the property of the double uniform linear arrays, i.e., the noise of arrays at different axis are uncorrelated, two noise free cross-correlation matrices are obtained theoretically. By using the property of shift-invariant of both array manifolds, first a cross-correlation matrix is calculated with the two maximum non-overlapping sub-arrays in x-axis and the maximum non-overlapping sub-array in y-axis, and a cross-correlation matrix is also computed with the two maximum non-overlapping sub-arrays in y-axis and the maximum non-overlapping sub-array in x-axis. ESPRIT algorithm is employed to generate two group DOAs with the two cross-correlation matrices, respectively. Then the two group DOAs are fused with variance fusion algorithm and higher precise DOA is obtained. Theoretical analysis shows the proposed algorithm is unbiased and can approach CRB at higher SNR. Monte Carlo simulations verify the effectiveness of the proposed method.