Underwater target time difference of arrival estimation using low-rank decomposition based generalized cross...

Area of Science:

• Underwater acoustics
• Signal processing
• Localization systems

Background:

• Acoustic waves are crucial for long-distance underwater detection and tracking.
• Passive localization using TDOA is geometrically constrained but sensitive to noise.
• Low signal-to-noise ratio significantly degrades TDOA estimation accuracy.

Purpose of the Study:

• To develop a robust TDOA estimation method for underwater acoustic localization.
• To enhance TDOA accuracy in low signal-to-noise ratio environments.
• To validate the proposed method with simulations and real-world experimental data.

Main Methods:

• Sub-band extraction from generalized cross correlation phase transform spectrum.
• Frequency sliding window technique for sub-band identification.
• Application of a low-rank decomposition framework (GoDec+) leveraging rank-one property.

Main Results:

• A novel robust TDOA estimation method based on sub-band extraction is proposed.
• The extracted sub-band matrix exhibits a rank-one property.
• The method demonstrates improved TDOA estimation accuracy in challenging acoustic conditions.

Conclusions:

• The proposed sub-band extraction method offers robust TDOA estimation for underwater localization.
• The rank-one property of sub-bands enables effective low-rank decomposition.
• The method shows promise for practical applications in marine acoustic systems.