Near-Field Multiple Sources Localization Via Sparse Reconstruction

The novel near-field sources localization methods based on sparse reconstruction are presented to estimate direction-of-arrival (DOA) and range efficiently. A special covariance matrix is first constructed, which decouples the DOA and range, and then we can obtain DOA estimations by solving the MUSIC-like weighting sparse minimization problem with l
2
-norm. After obtaining the estimated DOAs, the related range parameters can be obtained by the same sparse reconstruction method as the DOA estimate. Furthermore, an improved weighting sparse optimization problem based on l
1
-norm is proposed to enhance the estimator's robustness. Finally, several simulations illustrate the advantages of the proposed algorithms by comparison.