Microphone directivity-based single source point detection in multi-source direction-of-arrival estimation.

Area of Science:

• Acoustics
• Signal Processing
• Array Signal Processing

Background:

• Direction-of-arrival (DOA) estimation is crucial for source localization.
• Existing methods struggle with multi-source scenarios and acoustic challenges like noise and reflections.
• Single source point (SSP) detection is key to isolating individual sources in complex sound fields.

Purpose of the Study:

• To develop robust methods for single source point (SSP) detection in multi-source direction-of-arrival (DOA) estimation.
• To enhance the accuracy and stability of DOA estimation using ambisonic arrays.
• To overcome limitations of traditional methods in noisy and reverberant environments.

Main Methods:

• Proposed a phase-based SSP detection method utilizing virtual array signals with adjustable directional gain.
• Introduced a complementary SSP detection method based on directional gain relationships between microphones.
• Integrated both methods for enhanced SSP detection robustness.
• Employed kernel density estimation and peak search for final DOA estimation from detected SSPs.

Main Results:

• The proposed virtual array approach dynamically adjusts directional gain, improving signal strength and phase stability.
• Integration of phase-based and gain-based SSP detection methods leads to more robust performance.
• The combined method demonstrated superior DOA estimation accuracy in simulated and real-world tests compared to existing techniques.

Conclusions:

• The developed SSP detection techniques significantly improve multi-source DOA estimation accuracy.
• The virtual array concept offers advantages over physical arrays by mitigating reception blind regions.
• The integrated approach provides a robust solution for challenging acoustic environments, advancing sound source localization technology.