Two-Step 3D Microphone Array Fusion Algorithm to Enhance Sound Source Location Measurements

Area of Science:

• Acoustics and Signal Processing
• Computational Auditory Scene Analysis

Background:

• Sound source localization (SSL) in 3D indoor reverberant environments is challenging due to Angle-of-Arrival (AoA) variations.
• Existing methods struggle with accuracy dependent on source position and microphone array configuration.

Purpose of the Study:

• To develop and validate a novel two-step algorithm for microphone array fusion to enhance SSL accuracy.
• To generalize the algorithm for arbitrary room geometries and microphone array placements.

Main Methods:

• A two-step algorithm excluding low-confidence AoA microphone pairs was proposed.
• A generalized version was developed and tested across varied room dimensions and array positions.
• Performance was evaluated against conventional AoA estimation methods.

Main Results:

• The algorithm significantly improves SSL accuracy by mitigating unreliable AoA estimates.
• The generalized algorithm consistently reduces localization error by approximately 50% across diverse scenarios.
• Robustness was demonstrated across different room geometries and array placements.

Conclusions:

• The proposed algorithm offers a robust and practical solution for enhancing SSL in acoustically challenging environments.
• The generalized approach maintains significant error reduction regardless of environmental specifics.
• This work provides a valuable improvement over baseline SSL methods.