Passive Localization of Near-Field Sources Using Cumulant

Passive source localization is one of the issues in array signal processing fields. When some source lies in the Fresnel region of the array (i.e., near-field), the wavefront of the source must be characterized by both the direction-of-arrival (DOA) and range. To localize near-field sources, some classical far-field source localization methods are modified to estimate DOA and range at the costs of array aperture, which contains two hands: (i) quarter- wavelength constraint is imposed on the intersensor spacing and (ii) fewer sources in the near-field case are localized from an array of given sensor number than those of the far-field case. In addition, it is necessary to match two separately estimated electric angles (function of DOA and range) of multiple near-field sources, i.e., encountering parameter match problem. To solve these problems, this paper develops a new cumulant-based algorithm (CBA). The key points of this paper are: (i) based on the rank-reduction idea, the nonlinear optimization problem of two-dimensional parameters is decoupled to avoid both pairing parameters and two-dimensional search; (ii) unlike some classical near-field source localization algorithms, the quarter-wavelength element spacing constraint is not required in the proposed algorithm. Simulation results are presented to validate the performance of the proposed method.