Mode separation with one hydrophone in shallow water: A sparse Bayesian learning approach based on phase speed

Area of Science:

• Acoustics
• Oceanography
• Signal Processing

Background:

• Shallow water environments present challenges for acoustic signal analysis due to complex wave propagation.
• Separating acoustic modes is crucial for understanding underwater sound fields.
• Existing methods like time warping transformation struggle in environments with coexisting reflected and refracted modes.

Purpose of the Study:

• To develop a novel broadband mode separation technique for shallow water acoustics.
• To utilize sparse Bayesian learning (SBL) for enhanced mode separation.
• To provide a method that does not require prior knowledge of the seafloor properties.

Main Methods:

• Employing phase speed extracted from single hydrophone measurements.
• Utilizing the approximate modal dispersion relation to construct a dictionary matrix for SBL.
• Applying SBL to estimate sparse coefficients from multi-frequency pressure data.
• Retrieving separated normal modes using estimated coefficients and dictionary atoms.

Main Results:

• The proposed SBL-based method successfully separates broadband acoustic modes in shallow water.
• The technique is effective for both impulsive and known-form signals.
• Simulations show superior performance in environments with coexisting reflected and refracted modes compared to time warping transformation.

Conclusions:

• The developed approach offers a robust solution for mode separation in challenging shallow water acoustics.
• Sparse Bayesian learning provides an effective framework for acoustic mode separation.
• This method advances underwater acoustic signal processing capabilities without requiring bottom information.