Related Experiment Video
Updated: Aug 7, 2025
16:05
Using Micro-Electro-Mechanical Systems MEMS to Develop Diagnostic Tools
Published on: October 1, 2007
7.6K
MEMS piezoelectric resonant microphone array for lung sound classification.
Hai Liu1, Matin Barekatain1, Akash Roy1
1Department of Electrical and Computer Engineering, University of Southern California, Los Angeles, CA, United States of America.
Journal of Micromechanics and Microengineering : Structures, Devices, and Systems
|March 13, 2023
Summary
A new resonant microphone array (RMA) using piezoelectric microelectromechanical systems (MEMS) enhances lung sound analysis. This sensitive device improves wheezing detection and classification accuracy for both computer and wearable applications.
Area of Science:
- Engineering
- Biomedical Engineering
- Acoustics
Background:
- Lung sound auscultation is crucial for diagnosing respiratory conditions.
- Traditional microphones may not optimally capture subtle wheezing frequencies.
- Microelectromechanical systems (MEMS) offer miniaturization and high sensitivity for acoustic sensing.
Purpose of the Study:
- To develop and evaluate a highly sensitive piezoelectric MEMS resonant microphone array (RMA) for improved detection and classification of wheezing in lung sounds.
- To assess the RMA's performance compared to a traditional reference microphone.
Main Methods:
- Fabrication of an 8-microphone RMA with width-stepped cantilever resonators.
- Resonance frequencies were Mel-distributed between 230-630 Hz, covering the primary wheezing range.
- Characterization of microphone sensitivity and signal-to-noise ratios (SNRs) across relevant frequencies.
- Testing with deep learning (computer) and simple machine learning (wearable chipsets) algorithms.
Main Results:
- Microphone sensitivities ranged from 86 to 265 mV/Pa at resonance and 35 to 265 mV/Pa over 200-650 Hz.
- SNRs were high, between 86.6-98.0 dBA at resonance and 79-98 dBA over 200-650 Hz.
- Wheezing features were more distinguishable with the RMA, leading to higher automatic classification accuracy.
Conclusions:
- The piezoelectric MEMS RMA demonstrates superior performance for capturing and analyzing wheezing sounds.
- The enhanced signal distinguishability facilitates more accurate automated wheezing classification.
- This technology holds promise for advanced respiratory monitoring in both clinical and wearable settings.
