2025_programme: MAMBAT: An automated model-based framework for multiple animal acoustic tracking in challenging marine environments

• Day: June 16, Monday
    Location / Time: B. ERATO at 17:40-18:00
• Last minutes changes: -
• Session: 07. Inverse Problems in Acoustical Oceanography
  Organiser(s): Julien Bonnel, Stan Dosso
  Chairperson(s): Julien Bonnel, Stan Dosso
• Lecture: MAMBAT: An automated model-based framework for multiple animal acoustic tracking in challenging marine environments
  Paper ID: 2169
  Author(s): Pina Gruden, Eva-Marie Nosal, E. Elizabeth Henderson
  Presenter: Pina Gruden
  Abstract: Marine mammal localization and tracking plays an important role in passive acoustic monitoring which can be used to aid management and mitigation decisions. However, localizing and tracking can be challenging when multiple animals vocalize simultaneously, and sensors are far spaced. For this purpose we developed MAMBAT (Multiple-Animal Model-Based Acoustic Tracking) to fully automate the process of model-based passive acoustic methods for tracking multiple marine mammals. This framework uses ”track-before-localize” and “localize-then-track” strategies, which incorporate multi-target Bayesian tracking and model-based localization to account for depth-dependent sound speed profiles. MAMBAT was previously demonstrated on single and multiple sperm whale encounters from widely-spaced bottom-mounted sensors at the Atlantic Undersea Test and Evaluation Center in the Bahamas. Here, we present improvements to the framework and evaluate its application to more challenging scenarios (more animals, lower signal-to-noise-ratios) and different species from bottom-mounted sensors at the Pacific Missile Range Facility in Hawaii [work supported by the ONR Marine Mammals and Biology program].
• Corresponding author: Dr Pina Gruden
  Affiliation: Pina Gruden s.p.
  Country: Slovenia