Correlation-based passive localization: Linear system modeling and sparsity-aware optimization.

Area of Science:

• Acoustics
• Signal Processing
• Oceanography

Background:

• Matched field processing in underwater passive localization faces significant mismatch challenges.
• Data-derived replicas, particularly cross-correlation based ones, offer improved robustness but suffer from high sidelobes and require extensive frequency samples.
• Existing literature highlights these issues but lacks theoretical analysis.

Purpose of the Study:

• To theoretically analyze and address the high sidelobe and grating lobe challenges in cross-correlation based matched field processing.
• To develop a novel optimization approach for enhanced underwater passive localization.
• To extend the methodology for distributed sensor networks.

Main Methods:

• Revisiting conventional correlation-based matching by formulating the ambiguity surface as a least-norm solution of an underdetermined linear system.
• Recasting the localization problem as a sparsity-aware optimization problem to tackle sidelobes and measurement requirements.
• Validating the approach using simulated waveguide data, microphone data in air, and SWellEx-96 data.

Main Results:

• The proposed optimization perspective provides a theoretical understanding of sidelobes.
• The sparsity-aware approach effectively mitigates high sidelobes and reduces the need for extensive frequency samples.
• Demonstrated superior localization performance across diverse datasets and extended applicability to distributed sensor networks.

Conclusions:

• The novel sparsity-aware optimization framework offers a robust solution to longstanding challenges in matched field processing for passive localization.
• This approach significantly improves localization accuracy and efficiency in various acoustic environments.
• The extension to distributed sensor networks highlights its potential for advanced underwater surveillance and monitoring.