Method of Parameters Estimation for Mixed Sources Based on Sparse Reconstruction

In this letter, we present a parameters estimation algorithm based on the sparse signal reconstruction to cope with the mixed sources localization problem. First, the far-field sources parameters are estimated by the sparse reconstruction of the covariance matrix of the received signal. Then the covariance separation technique is exploited to separate the near-field and the far-field sources. Finally, the near-field source parameters are estimated by the symmetric property of the uniform linear array geometry and the sparse signal reconstruction. Two-dimensional peak searching and the matching of the near field sources parameters as well as the construction of the high-order cumulants are avoided, which reduce the computational complexity. The simulation results show that the spatial resolutions and the estimation accuracy of the mixed sources parameters are higher than those based on subspace.