This paper proposes the use of gated feedback gated recurrent unit network (GFGRU), a learning-based sparse estimation algorithm, for multiple source localization in the direct arrival zone of the deep ocean. The GFGRU, trained on sound field replicas of a single source generated by an acoustic propagation model, is used to estimate the ranges and depths of multiple sources without knowing the number of sources. The performance of GFGRU is compared to the Bartlett processor, feedforward neural network (FNN), and sparse Bayesian Learning (SBL) algorithm. Simulations indicate that GFGRU behaves similarly to SBL and offers modest localization performance improvement over the Bartlett and FNN in the presence of array tilt mismatch. The results of real data from the South China Sea also verify the robustness of the proposed GFGRU using a 105 m-aperture vertical array in the deep ocean.

The structure of a sound speed profile (SSP) in deep water causes refraction of sound rays and Convergence Zones (CZs) of high intensity where the rays focus at shallow depth. Study of sound field characteristics in the CZs has always been the focus of deep-water acoustics research. Many studies have been conducted on sound propagation in different parts of the oceans with different environments and, in this paper, the range and width of CZ is analyzed in the East Indian Ocean (EIO) and the South China Sea (SCS). Through the experimental data collected in different seasons with the propagation conditions change in the EIO and the SCS, we observe that the SSPs in different marine environments have a significant impact on the CZs of deep water. The sound channel mixing layer and isothermal layer have great effect on the CZ ranges. The water depths in the two experimental areas are similar, the range of the first CZ in the EIO is 7–8 km farther than that in the SCS, and the width of the CZs in the EIO is about 2–3 km narrower than that in the SCS. The surface mixed layer and the thermocline affect the CZ width but has little effect on the CZ range when the sound speed at the source and the bottom are practically the same. As the propagation conditions change along the seasons in the EIO, the range of the first CZ is almost the same, but the width of the CZs in the summer is about 2 km narrower than that in the spring. The water depth affects the CZ width but has little effect on the CZ range if the CZs can be formed. The different CZ characteristics between EIO and SCS are explained by both theoretical calculation and numerical simulation. The influence of the SSP structure and water depth on the CZ range are analyzed and the corresponding mechanism is explained. The research results are of great significance for underwater acoustic detection in deep sea.

In the direct arrival zone of the deep ocean, the multi-path time delays have been used for acoustic source localization. One of the challenges in conventional localization methods is to artificially determine which paths the extracted delays belong to. A convolutional neural network, taking the autocorrelation functions as the input feature directly, is proposed for source localization to avoid the path determination procedure. Since some multi-path arrivals may not be visible due to absorption in the bottom of the ocean, a data augmentation method based on a ray propagation model is proposed. Tests on simulated and real data validate the method.

Deep-learning source localization methods trained on synthetic replicas calculated by acoustic propagation models (referred to as model-based source localization methods via deep learning) suffer from environmental mismatches between replicas and measurements, which degrade the localization performance. This article introduces deep transfer learning (DTL), specifically the unsupervised domain adaptation (UDA) and fine-tuning methods for source localization using ships of opportunity (SOO) to alleviate the influence of mismatched replicas. First, the range estimation network is trained using simulated replicas and unlabeled SOO data. Subsequently, the parameters of the ranging network are transferred to a depth estimation network and fine-tuned accordingly. Simulation results demonstrate that the proposed method (UDA-DTL) exhibits greater robustness to array tilt and sound speed profile mismatches in the direct arrival zone compared to the Bartlett processor and the ResNet estimator. Experimental data from the South China Sea also verified the validity of the proposed UDA-DTL estimator. The results demonstrate that supplementing model-based supervised learning with unlabeled SOO, the impact of unknown uncertainty on the localization performance is alleviated. This study highlights the potential to utilize unlabeled Big Data in applications of underwater source localization.