Facilitating at-home spirometry with millimeter-wave devices

Abstract

An integrated system and associated methodology allow performing at-home spirometry tests using smart devices which leverage the built-in millimeter-wave (mmWave) technology. Implementations leverage deep learning with some embodiments including a combination of mmWave signal processing and CNN-LSTM (Convolutional Neural Network-Long Short-Term Memory Network) architecture. Smartphone devices are transformed into reliable at-home spirometers by having a user hold a device in front of their mouth, inhale their full lung volume, and forcibly exhale until the entire volume is expelled, as in typical spirometry tests. Airflow on the device surface creates tiny vibrations which directly affect the phase of the reflected mmWave signal from nearby objects. Stronger airflow yields larger vibration and higher phase change. The technology analyzes tiny vibrations created by airflow on the device surface and combines wireless signal processing with deep learning. The resulting low-cost, contactless method of lung function monitoring is not affected by noise and motion and provides all key spirometry indicators.