A covariance fitting approach for correlated acoustic source mapping

Area of Science:

• Acoustics
• Signal Processing
• Aeroacoustics

Background:

• Microphone arrays are crucial for aeroacoustic noise source localization and power estimation.
• Traditional delay-and-sum (DAS) beamformers have limitations in resolution and sidelobe levels.
• Existing deconvolution methods like DAMAS struggle with correlated sources or are computationally intensive.

Purpose of the Study:

• To introduce a computationally efficient deconvolution approach for mapping acoustic sources, particularly correlated ones.
• To address the limitations of existing methods in handling correlated sources in aeroacoustic measurements.
• To present a novel covariance fitting approach named MACS.

Main Methods:

• Developed a covariance fitting approach for Mapping of Acoustic Correlated Sources (MACS).
• MACS utilizes convex optimization and sparsity to minimize a quadratic cost function cyclically.
• The method is guaranteed to converge at least locally, offering improved computational efficiency.

Main Results:

• MACS effectively handles uncorrelated, partially correlated, and coherent acoustic sources.
• Demonstrated performance through simulations and experimental data.
• Achieved reasonably low computational complexity compared to existing methods for correlated sources.

Conclusions:

• MACS provides a practical and computationally efficient solution for acoustic source mapping with complex source correlations.
• The method enhances the accuracy of noise source localization and power estimation in aeroacoustics.
• MACS represents a significant advancement over traditional beamforming and deconvolution techniques for correlated sources.