Grid-free compressive mode extraction.

Area of Science:

• Ocean acoustics
• Signal processing
• Geophysics

Background:

• Extracting acoustic normal modes from ocean waveguide data is crucial for understanding sound propagation.
• Conventional methods often struggle with issues like wavenumber search grid discretization and require a priori environmental knowledge.
• Vertical Line Arrays (VLA) provide depth-resolved acoustic measurements essential for mode analysis.

Purpose of the Study:

• To present a novel grid-free compressive sensing (CS) method for extracting acoustic normal modes from VLA data.
• To improve the accuracy and robustness of mode horizontal wavenumber and shape estimation.
• To demonstrate the method's effectiveness without requiring a priori environmental knowledge or the mode orthogonality condition.

Main Methods:

• Utilizing a grid-free compressive sensing approach based on group total-variation norm minimization.
• Applying sparse representation of waveguide propagation using modes at discrete horizontal wavenumbers.
• Jointly processing multiple sensor data from a VLA at various depths and ranges.

Main Results:

• Successfully extracted mode wavenumbers and shapes from simulated and experimental (SWellEx-96) data.
• Demonstrated improved estimation performance through joint processing of multiple sensor data compared to single sensor processing.
• Showcased the method's ability to function effectively with partial water column data and without mode orthogonality assumptions.

Conclusions:

• The grid-free CS method offers a robust and efficient solution for acoustic normal mode extraction in ocean waveguides.
• Joint processing of VLA data significantly enhances the accuracy of mode parameter estimation.
• This approach advances underwater acoustic signal processing capabilities, particularly in scenarios with limited prior information.