Two-dimensional direction estimation of coherent signals with two parallel uniform linear arrays

In this paper, the problem of estimating two-dimensional (2D) direction-of-arrivals (DOAs) of coherent signals is investigated, and a new computationally efficient subspace-based method is proposed for the multiple coherent narrowband signals impinging on a planar sensor array formed by two uniform linear arrays (ULAs), where the array geometry and shift invariance property of the ULA are exploited to decorrelate the coherency of incident signals, and the cross-correlations between two ULAs are used to mitigate the effect of additive noise in null space estimation, while the pair-matching of the estimated azimuth and elevation angles and computationally burdensome eigendecomposition are avoided. The asymptotic mean-square-error (MSE) expressions of the estimated azimuth and elevation angles are derived explicitly. The effectiveness of the proposed method and the theoretical analysis are substantiated through numerical examples.