Study On DOA Estimation Of Coherent Signals Method With L-Shaped Array

Two-dimensional(2-D) direction-of-arrival(DOA) estimation of multiple incident signals has attracted considerable attention in array signal processing fields. Recently, since L-shaped arrays have some merits in coverage area and performance, there has been growing interest in developing 2-D DOA estimation with L-shaped arrays. However, many methods may fail to work when signals are coherent due to multipath propagation in real environments. Thus the de-correlation algorithm for DOA estimation of coherent signals with L-shaped array is mainly researched. The details are described as follows:(1) A new covariance matrix is constructed for two-dimensional(2-D) direction of arrival(DOA) estimation of narrowband coherent signals with L-shaped array. The covariance matrix is constructed by exploiting the forward/backward matrix of cross-correlations, and the constructed differencing matrix by auto-correlations where the contribution of the additive noise is mitigated. Due to the increased averaging for the same array aperture, the degree of de-correlation further increases compared with the previous methods. Moreover, a modified pair-matching method is presented by minimizing a constructed cost function without the noise term. Simulation results verify that the proposed method provides better robustness than conventional methods.(2) A modified correlation vector reconstruction method(MCVM) is proposed for DOA estimation of coherent signal with L-shaped array.The method is based on the theory of vectors reconstruction. By employing the array characteristics that the cross-correlation matrix is not affected by the additive noise, we use each column vector of the cross-correlation matrix as the forward vector, and the backward vector is obtained by reversing the order of the complex conjugated elements of the forward one. All forward and backward vectors are used to reconstruct the signal covariance matrix from which extract the signal subspace. The proposed method enlargers array aperture. Simulation results verify that the proposed method has excellent performance compared with the traditional algorithm.(3) A novel algorithm is proposed based on singular value decomposition(SVD) vector reconstruction for two-dimensional(2-D) direction of arrival(DOA) estimation of narrowband coherent signals with L-shaped array. The proposed method performs SVD to cross-correlation matrix which is not square matrix and utilizes mutually independent singular vectors corresponding to the signal source subspace to construct the covariance matrix which enlarges the array effective aperture. Moreover, a modified pair-matching method of estimated elevation and azimuth angles is presented by minimizing the constructed cost function without the noise term. Simulation results verify the effectiveness of the proposed algorithm and the proposed algorithm can resolve more signals.