2025_programme: Underwater Passive Source Localization based on a Data Derived Replica Library

• Day: June 16, Monday
    Location / Time: B. ERATO at 17:20-17:40
• Last minutes changes: -
• Session: 07. Inverse Problems in Acoustical Oceanography
  Organiser(s): Julien Bonnel, Stan Dosso
  Chairperson(s): Julien Bonnel, Stan Dosso
• Lecture: Underwater Passive Source Localization based on a Data Derived Replica Library
  Paper ID: 2273
  Author(s): Baptiste Menetrier, Gnouregma Bazile Kinda, Valérie Labat, Samuel Pinson, Abdel Boudraa
  Presenter: Baptiste Menetrier
  Abstract: Passive localization of underwater acoustic source using a sparse sensor network is a challenging task due to the complex nature of the ocean environment. Matched Field Processing (MFP) techniques have drawn considerable attention in the literature as a potential solution to this challenge. Generally speaking, this technique can be viewed as a beamformer which matches the measured data with dictionary replicas to estimate the source location. However, its successful implementation in real-life operating conditions remains limited due to sensitivity to environmental mismatch. In the case of low-frequency signals, such as radiated ship noise, uncertainties in waveguide properties such as bathymetric features or seabed parameters can lead to significant localization errors.\n\nTo address these challenges, MFP methods based on data-derived replicas have been recently explored. To overcome the limitations of cross-correlation-based data-derived replicas, this paper investigates the potential of Relative Transfer Function (RTF) vectors as localization features in an MFP approach. The distance between the RTF vector associated with an unknown radiator and the library vectors is evaluated using the so called Hermitian angle. The key idea of the method is to exploit the environmental diversity to construct replicas that smoothly map to source locations.\n\nThe potential of this approach is evaluated through simulations. A synthetic dataset is generated using the KRAKEN normal mode propagation code in a deep water environment and the ultra low frequency domain. Propagated time series are synthesized using a ship noise model as the source signal, and performance is assessed across different noise scenarios, including additive white Gaussian noise and interference from propagated signals. Results demonstrate the feasibility of the proposed method.
    Download the full paper
  This paper is a candidate for the "Prof. Leif Bjørnø Best Student Paper Award (for students under 35)"
• Corresponding author: Mr Baptiste Menetrier
  Affiliation: Arts et Métiers Institute of Technology, École navale, IRENav
  Country: France