A High-Resolution Time Reversal Method for Target Localization in Reverberant Environments.

Area of Science:

• Acoustics
• Signal Processing
• Machine Learning

Background:

• Reverberation significantly degrades the performance of sound source localization (SSL) methods.
• Broadband low-frequency signals are prevalent in real-world acoustic environments.
• Existing SSL techniques struggle with multipath effects caused by reverberation.

Purpose of the Study:

• To develop a robust and high-resolution SSL method for broadband low-frequency signals in reverberant conditions.
• To leverage the strengths of time reversal (TR) and sparse Bayesian learning (SBL) for improved localization accuracy.
• To address the challenges posed by multipath propagation and environmental noise.

Main Methods:

• Investigated sound propagation models in reverberant environments.
• Employed time reversal (TR) to achieve adaptive focusing and mitigate multipath effects.
• Utilized sparse Bayesian learning (SBL) for accurate wave direction estimation and source localization.
• Validated the proposed method through extensive simulations and experiments in a reverberation chamber.

Main Results:

• The proposed TR-SBL method demonstrated satisfactory spatial resolution in localizing broadband low-frequency sound sources.
• Performance was evaluated across various parameters including bandwidth, microphone array size, SNR, and reverberation time.
• The method proved robust under off-grid conditions and in challenging reverberant environments.

Conclusions:

• The developed SSL method effectively overcomes reverberation and multipath interference.
• The integration of TR and SBL offers a powerful approach for accurate sound source localization.
• The proposed technique shows significant promise for practical applications in acoustically complex environments.