Deep-learning source localization using multi-frequency magnitude-only data

• Main Authors: Gerstoft, P. (Author), Gong, Z. (Author), Li, Z. (Author), Niu, H. (Author), Ozanich, E. (Author), Wang, H. (Author)
• Format: Article
• Language: English
• Published: Acoustical Society of America 2019
• Subjects: Acoustic fields; Acoustic propagation; article; Backpropagation; Broadband acoustic source; China; deep learning; Deep learning; Learning approach; Learning sources; Magnitude only data; simulation; sound; Source localization; Two-step training; Uncertain environments; uncertainty; Uncertainty analysis; Yellow Sea
• Online Access: View Fulltext in Publisher

 A deep learning approach based on big data is proposed to locate broadband acoustic sources using a single hydrophone in ocean waveguides with uncertain bottom parameters. Several 50-layer residual neural networks, trained on a huge number of sound field replicas generated by an acoustic propagation model, are used to handle the bottom uncertainty in source localization. A two-step training strategy is presented to improve the training of the deep models. First, the range is discretized in a coarse (5 km) grid. Subsequently, the source range within the selected interval and source depth are discretized on a finer (0.1 km and 2 m) grid. The deep learning methods were demonstrated for simulated magnitude-only multi-frequency data in uncertain environments. Experimental data from the China Yellow Sea also validated the approach. © 2019 Acoustical Society of America.