Matrix completion based direction-of-arrival estimation in nonuniform noise

It is known that eigenstructure-based direction-of-arrival (DOA) estimation algorithms are vulnerable to nonuniform noise. In order to tackle this problem, recently we proposed an approach for joint estimasion of the signal subspace and noise covariance matrix. However, an iterative procedure is involved in this method. This motivates us to present an new approach which is free of iteration in this paper. More precisely, the problem of noise-free covariance matrix estimation is first formulated as matrix completion. The signal and noise subspaces are then achieved by eigendecomposing the noise-free covariance matrix estimate and, therefore, traditional subspace-based DOA estimation algorithms can be applied directly. Numerical simulation results are provided to illustrate the effectiveness of the proposed method.