Sound speed estimation and source localization with linearization and particle filtering.

Area of Science:

• Ocean acoustics
• Geophysical signal processing

Background:

• Accurate estimation of underwater source location and sound speed is crucial for various oceanographic applications.
• Traditional methods face challenges with noise and complex environmental conditions.

Purpose of the Study:

• To develop and validate an efficient linearization method for estimating source location and sound speed in the water column.
• To assess the method's accuracy and robustness under varying noise levels.

Main Methods:

• A novel linearization approach using empirical orthogonal function coefficients in the Jacobian matrix.
• Testing the inversion technique with synthetic arrival times and Gaussian error distributions.
• Employing particle filtering for arrival time estimation from experimental data (Shallow Water 06).
• Propagating particle filter-derived probability density functions through the inversion process.

Main Results:

• The linearization method demonstrates efficiency and accuracy with few iterations on synthetic data.
• Probability densities of estimates were successfully calculated for different noise levels.
• Particle filtering effectively estimated arrival times from experimental signals.
• Inversion estimates for source location and sound speed are consistent with literature values and fast simulated annealing results.

Conclusions:

• The proposed linearization method provides accurate and efficient estimation of underwater source location and sound speed.
• Particle filtering enhances arrival time estimation, leading to reliable inversion results.
• The method shows robustness and consistency, making it a valuable tool for ocean acoustic studies.