Abstract
This paper proposes a new cumulant-based algorithm to jointly estimate four-dimensional (4D) source parameters of multiple near-field narrowband sources. Firstly, this approach proposes a new cross-array, and constructs five high-dimensional Toeplitz matrices using the fourth-order cumulants of some properly chosen sensor outputs; secondly, it forms a parallel factor (PARAFAC) model in the cumulant domain using these matrices, and analyzes the unique low-rank decomposition of this model; thirdly, it jointly estimates the frequency, two-dimensional (2D) directions-of-arrival (DOAs), and range of each near-field source from the matrices via the low-rank three-way array (TWA) decomposition. In comparison with some available methods, the proposed algorithm, which efficiently makes use of the array aperture, can localize N -3 sources using N sensors. In addition, it requires neither pairing parameters nor multidimensional search. Simulation results are presented to validate the performance of the proposed method.
Highlights
Estimation of directions-of-arrival (DOAs) has received a significant amount of attention over the last several decades
This paper proposes a new cumulant-based algorithm to jointly estimate four-dimensional (4D) source parameters of multiple near-field narrowband sources
The above-mentioned analyses show that the main difficulties of near-field source localization problem consist of: (i) avoiding multidimensional search which results in extremely demanding computational complexity; (ii) reducing the loss of the array aperture; (iii) pairing source parameters so as to localize the near-field sources accurately
Summary
Estimation of directions-of-arrival (DOAs) has received a significant amount of attention over the last several decades. To solve near-field source localization problem, many algorithms were addressed, such as the ML method [5], the 2D MUSIC methods [6,7,8,9], the linear prediction methods [10, 11], and the ESPRIT-like methods [12,13,14,15]. EURASIP Journal on Advances in Signal Processing method, which suffers a heavy loss of the array aperture, can localize not more than (1/4)(N − 5) sources using N sensors It requires a quadratic phase transform algorithm to pair the separately estimated parameters. The above-mentioned analyses show that the main difficulties of near-field source localization problem consist of: (i) avoiding multidimensional search which results in extremely demanding computational complexity; (ii) reducing the loss of the array aperture; (iii) pairing source parameters (i.e., frequency, azimuth, elevation, and range) so as to localize the near-field sources accurately.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
