A Sparse Array DOA Estimation Approach Based on Matrix Completion

In order to improve the performance of direction of arrival (DOA) estimation for sparse array, this paper applies low-rank matrix completion theory to DOA estimation and proposes an improved matrix completion model and optimization solution. The proposed method utilizes the Hermitian and Toeplitz structural information of the observation matrix as a priori information. We split the augmented Lagrangian function and minimize one of the subproblems by Dykstra alternating projection which makes the completion matrices keep the Toeplitz structure and accomplishes the denoising of the known data. The simulation results show that the proposed method can effectively realize the reconstruction of sparse array, the performance of DOA estimation is excellent, and it can be applied to coherent sources.