Contactless Respiration Monitoring Using Ultrasound Signal

Respiratory monitoring is of great importance in various applications, including health monitoring, early disease detection, etc. Traditional methods often rely on dedicated devices, which can be intrusive and inconvenient. In this paper, we propose UltraResP, a non-invasive and contactless respiratory monitoring system that utilizes a single pair of smartphoneintegrated speaker and microphone. UltraResP periodically emits Frequency Modulated Continuous Wave (FMCW) chirps and captures the reflected signals from the human body. We introduce a transmission time-based strategy to accurately identify the target range bin where respiration-induced signals are concentrated, enabling robust recovery of the respiration waveform. Furthermore, we design a deep learning model to refine the extracted waveform, enhancing estimation accuracy. Extensive experiments under various conditions demonstrate the effectiveness and reliability of UltraResP in real-world scenarios.