A new near-field source localization algorithm based on generalized ESPRIT

By constructing a novel symmetric subarrays structure, a new 3-D near-field source localization algorithm is proposed. The new method can be divided into three important parts. The first part is the elevation angle estimation based on two generalized ESPRIT algorithm by a vertical uniform linear array, which are spectral search-based generalized ESPRIT and search-free generalized ESPRIT. And the second part is the range estimation based on 1-D MUSIC or MVDR spectral search. The third part is azimuth angle estimation by a single vector hydrophone lie in the center of this vertical array. We form the MVDR spectrum for azimuth angle estimation based on each pair of estimated elevation angle and range. For sources, in this new 3-D near-field localization algorithm, a total of times 1-D search is required base on spectral search-based generalized ESPRIT, and times 1-D search is required base on search-free generalized ESPRIT. Both algorithms do not require parameter pairing or high-order statistics computation, and can estimate 3-D near-field parameter only by linear array instead of geometry array.