Particle filtering for dispersion curve tracking in ocean acoustics

Area of Science:

• Ocean acoustics
• Signal processing
• Geophysics

Background:

• Extracting modal dispersion from underwater acoustic signals is crucial for geoacoustic inversion and source localization.
• Traditional time-frequency analysis methods often lack accuracy and uncertainty quantification for dispersion characteristics.

Discussion:

• A novel particle filtering method is presented for robust dispersion curve extraction from broadband acoustic spectrograms.
• The method effectively estimates waveguide dispersion characteristics, outperforming conventional time-frequency analysis techniques.
• It provides valuable uncertainty information on modal arrival times, a feature often missing in existing approaches.

Key Insights:

• The particle filtering approach yields highly accurate modal dispersion estimates, closely matching numerically calculated curves.
• This method enhances the precision of underwater acoustic data analysis for geophysical applications.
• Uncertainty quantification in modal arrival times is a significant advancement for robust inversion models.

Outlook:

• Future research could involve applying this method to real-world oceanographic data for validation.
• The technique holds potential for improving the performance of autonomous underwater vehicles in navigation and environmental monitoring.
• Further development may extend its applicability to other wave propagation phenomena in complex media.