Large-region acoustic source mapping using a movable array and sparse covariance fitting.

Area of Science:

• Acoustics
• Signal Processing
• Computational Engineering

Background:

• Accurate acoustic source mapping is crucial for city-scale noise monitoring.
• Existing methods using single-position measurements or large arrays are insufficient or too costly.
• A novel multiple-position measurement scheme is needed for efficient large-area acoustic analysis.

Purpose of the Study:

• To develop an efficient and cost-effective method for large-region acoustic source mapping.
• To improve the quality of acoustic source maps and reduce sound pressure level errors.
• To present a sparse-constrained multiple-position vectorized covariance matrix fitting approach.

Main Methods:

• Utilizing a multiple-position measurement scheme with a movable small-sized array.
• Estimating the sample covariance matrix of an incoherent virtual array via the Khatri-Rao (KR) product.
• Applying a sparse-constrained reconstruction algorithm to recover source powers from a vectorized covariance matrix model.

Main Results:

• The proposed method generated significantly cleaner acoustic source maps.
• Lower sound pressure level errors were achieved compared to beamforming and previous sparse methods.
• The approach was validated using simulated and measured data across different region sizes.

Conclusions:

• The sparse-constrained multiple-position vectorized covariance matrix fitting approach offers superior performance for large-region acoustic source mapping.
• This method provides a more accurate and cost-effective solution for city-scale noise monitoring.
• The findings suggest a promising direction for advanced acoustic imaging and analysis.