- 1Université de Toulon, Aix Marseille Univ, CNRS, DYNI, LIS, Toulon, France
- 2Centre International d'Intelligence Artificielle en Acoustique Naturelle
Access to the 3D tracks of sperm whales is required to assess the impact of anthropophony on their diving behavior. Passive acoustic and the time difference of arrival (TDoAs) are methods commonly used to localize the source of the clicks, although require a large number of hydrophone arrays [4] or are imprecise. While azimuths and elevations are readily assessed via TDoAs with a single antenna [1,2,5], radius estimation is challenging for small arrays with respect to the Cramer Rao Bound [1,5]. We therefore exploit other relationships between TDoAs, radius and frequency difference of arrival (FDoAs) [3]. In order to solve such systems, deep learning models have been trained with supervised learning and generated dataset to enable radius estimation using a single relationship while combining observables of different natures (i.e. TDoAs and FDoAs). The model achieves a 10% error in radius estimation in most cases. This neural network model has also been tested on fieldwork data with small arrays and gives promising biodiversity monitoring.
[1] Bénard-Caudal, F. Giraudet, P. Glotin, H. (2010) Whale 3D monitoring using astrophysic NEMO ONDE two meters wide platform with state optimal filtering by Rao-Blackwell Monte Carlo data association, Applied acoustics 71 11 994-999
[2] Giraudet, P., Glotin, H. (2006). Real-time 3D tracking of whales by echo-robust precise TDOA estimates with a widely-spaced hydrophone array. Applied Acoustics, 67(11-12), 1106-1117.
[3] Glotin, H. Mishchenko, A. Giraudet, P. (2015) Contraintes conjointes de différences temporelles et effet Doppler multibandes pour la séparation, caractérisation et localisation de sources sonores par acoustique passive. Patent FR14 54539.2015
[4] Nosal, E. M., & Frazer, L. N. (2006). Track of a sperm whale from delays between direct and surface-reflected clicks. Applied Acoustics, 67(11-12), 1187-1201.
[5] Zimmer, W. M. (2013). Range estimation of cetaceans with compact volumetric arrays. The Journal of the Acoustical Society of America, 134(3), 2610-2618
How to cite: Buisson, M.-L. and Glotin, H.: Subsea source range estimation based on Neural Network, One Ocean Science Congress 2025, Nice, France, 3–6 Jun 2025, OOS2025-1119, https://doi.org/10.5194/oos2025-1119, 2025.
