Smartphone Acoustic Sensing for Contactless Respiration Monitoring and Gesture-Based Authentication
This paper presents a smartphone-based sonar-sensing system for contactless monitoring of respiration and gesture-based user activity, using a single commercial phone with no hardware modification. A 20 kHz audio signal is transmitted from the phone’s speaker and reflections are captured by its microphone; motion-induced Doppler shifts are recovered using an envelope detection method that removes the need for a reference copy of the transmitted signal and improves robustness to transmitter frequency drift. Actuator experiments at a target distance around 50 cm demonstrated motion sensitivity at two conditions, 6 mm at 0.15 Hz and 2 mm at 0.6 Hz, with a measured signal-to-noise ratio of approximately 21.7 dB. Respiration rate detected by the system was compared against a pulse oximeter across two subjects, with an error within 0.1 Hz of the reference. Gesture-based interactions (tapping, scrolling, typing, talking) were analyzed using short-time Fourier transform features as a first step toward device authentication. Distance and direction estimation are outside the scope of this work and are identified as directions for future development.