A new method for 3D localization using a hybrid array in shallow ocean with non-Gaussian noise

In this paper, a new method is presented for three dimensional (3D) localization of narrow band acoustic sources in shallow ocean with non-Gaussian noise. The noise is modeled as a Gaussian mixture with a small number of components. The new algorithm consists of an initialization step to obtain rough estimates of the unknown parameters followed by an update step in which the initial estimates are updated sequentially using a novel approach based on the space alternating generalized expectation maximization (SAGE) algorithm. A hybrid array composed of one horizontal subarray and one vertical subarray is used. The vertical subarray data is used for range-depth estimation. The horizontal subarray data and the range-depth estimates are used for bearing estimation. The proposed method is computationally simpler than existing methods because it replaces the conventional 3D search by a 2D search and a small number of 1D searches. Simulation results are presented to show that the root-mean-square (RMS) errors of estimates obtained by the new method are significantly lower than the RMS errors of MUSIC estimator. The RMS errors of the proposed method are also compared with the CRB