An SOS-Based Algorithm for Localization of Multiple Near-Field Sources Using Uniform Circular Array

Abstract

We present a second-order statistics (SOS)-based algorithm for three-dimensional (3-D) localization of multiple near-field sources using an uniform circular array (UCA). Based on the centro-symmetric property of UCA steering vectors, a particular diagonal matrix, which only includes two-dimensional (2-D) direction of arrival (DOA) is constructed by utilizing the rotational relationship of the different UCA outputs. By exploiting the subspace of an SOS for UCA output, an estimation of signal parameters via rotational invariance techniques (ESPRIT)-like and a one-dimensional MUSIC methods are performed to obtain the 2-D DOA and range spatial spectrum, respectively. The proposed SOS-based algorithm does not require high-order statistics computation and also does not require additional eigenvalue decomposition for range estimation. Simulation results demonstrate the superior performance of the proposed algorithm compared with a TSMUSIC method and a phase-based method.