Mixed sources localisation using a sparse representation of cumulant vectors

In this study, a new mixed near-field and far-field sources localisation algorithm based on sparse signal recovery is addressed. In this scheme, two special cumulant vectors are constructed successively, the first one is used to obtain the azimuth estimations of all the incoming signals, and the second one is used to distinguish the mixed sources as well as estimate the range related to the near-field sources. The reweighted ℓ₁-norm minimisation with one iteration is utilised for sparse signal recovery. In the recovery process, the authors propose to select the regularisation parameter by a special case of 2-fold cross-validation. The simulation results demonstrate the effectiveness and the efficiency of the proposed algorithm.