Estimation DOAs of the coherent sources based on SVD

An improved Method of DOA estimation in the presence of coherent signals is presented. First, using the eigenvector corresponding to the largest eigenvalue obtained from the received data covariance matrix, the presented method rebuilds a matrix, with which the auto-covariance matrix can be obtained. Then, we reconstruct a matrix using the equal of the obtained auto-covariance matrix and its complex-conjugation inversion. Thus, noise subspace can be gained by singular value decomposition. Afterwards, the direction parameters are estimated based on the MUSIC method. In comparison with the conventional de-correlation algorithms, the method offers better estimation performance for coherent sources, especially at low SNR and small snapshots. Simulation experiments demonstrate the effectiveness of the proposed method for estimation DOAs of the coherent sources.