2023_programme: Machine learning techniques for source ranging and sediment classification in the ocean
- Session: 16. Trends and Advances in Array Signal Processing
Organiser(s): Angeliki Xenaki, Peter Gerstoft and Eliza Michalopoulou
- Lecture: Machine learning techniques for source ranging and sediment classification in the ocean [invited]
Paper ID: 1898
Author(s): Frederick Christina, Michalopoulou Zoi-Heleni
Presenter: Michalopoulou Zoi-Heleni
Abstract: In this talk we discuss machine learning approaches for inverse problems in underwater acoustics, where the goal is to recover detailed characteristics of the seafloor from measured data generated from SONAR systems as well as localize the sound emitting source. The key to successful inversion is the use of accurate forward modeling that captures the dependence of sound propagation on seafloor properties, such as sediment type, roughness, and thickness of layers. To solve the inverse problem, machine learning strategies applied to a reference library of acoustic templates can be used to estimate the seafloor parameters that describe the full domain. Workshop ’97 data are employed for seabed classification and source range estimation. The data are acoustic fields computed at vertically separated receivers for various ranges and different environments. Gaussian processes are applied for denoising the data and predicting the field at virtual receivers, sampling the water column densely within the array aperture. The enhanced fields are used in combination with machine learning to map the signals to one of 15 sediment-range classes (corresponding to three environments and five ranges). The classification results after using Gaussian processes for denoising are superior to those when noisy workshop data are employed in terms of both classification performance and training requirements.
- Corresponding author: Prof Zoi-Heleni Michalopoulou
Affiliation: Department of Mathematical Sciences, New Jersey Institute of Technology
Country: United States
e-mail:
