Improving AoA Localization Accuracy in Wireless Acoustic Sensor Networks with Angular Probability Density Functions.

Area of Science:

• Electrical Engineering
• Acoustics
• Signal Processing

Background:

• Energy-efficient electronics enable wireless acoustic sensor networks (WASNs).
• WASNs can detect and locate sound events for surveillance and assisted living.
• Linear MEMS microphone arrays provide angular information for sound source localization.

Purpose of the Study:

• To develop and evaluate a novel sound event localization algorithm for WASNs.
• To compare the accuracy of a new matching algorithm with traditional triangulation methods.
• To investigate computationally efficient angle-of-arrival (AOA) calculation methods.

Main Methods:

• Utilized low-power linear MEMS microphone arrays for angular data acquisition.
• Developed a localization approach using angular probability density functions and a matching algorithm.
• Investigated two delay-based AOA calculation methods.
• Experimentally evaluated algorithms in a controlled room environment.

Main Results:

• The proposed matching algorithm demonstrated superior accuracy over common triangulation.
• An accuracy improvement of up to 114% was achieved when localizing white noise.
• Both white noise and vocal sounds were successfully localized.

Conclusions:

• The novel matching algorithm offers enhanced accuracy for sound event localization in WASNs.
• This approach provides a more precise alternative to traditional triangulation methods.
• The findings support the deployment of WASNs for accurate acoustic event detection.