Statistical Acoustic Sensing For Real-Time Respiration Monitoring and Presence Detection

In this demo, we present an all-in-one real-time system for breathing monitoring and presence detection using statistical acoustic sensing. By applying Auto-Correlation Function (ACF) to the Channel Frequency Response (CFR), our system captures both motion statistics and breathing rates. We devise novel weight combining schemes to enhance the SNR of the weak sensing signals. We then enable human presence detection by integrating both motion statistics and breathing rate as vital indicators. Our system operates using a single microphone without relying on a bulky microphone array. Our demo functions in real-time and supports any device that is equipped with a commodity microphone and speaker. Our demo can be accessed through https://youtu.be/1bxpXNwHGv0.