Parameters estimation of mixed sources based on sparse reconstruction

A new parameters estimation approach for mixed sources is proposed in this paper, which is under the sparse signal reconstruction framework. By reconstructing of covariance matrix, directions of arrival (DOAs) in the far-field part were estimated with an l ₁ -svd-like method directly. Then the subspace difference method was adopted to obtain the near-field covariance matrix, and the relative DOAs can be also estimated with the sparse signal recovery. Finally, the range parameters of near-field sources were estimated via the near-field covariance matrix reconstruction. Neither spectral research nor parameters match exist in our method. This method can distinguish the mixed sources better, and give an improved accuracy than the existing methods. Simulation results verify the performance of our method.