Feasibility study of practical vital sign detection using millimeter-wave radios

The importance of vital sign detection is self-evident in the mobile health domain. Recent work has shown that one can use RF or WiFi signals for respiration and heartbeat detection in a non-contact manner and thus improve its usability compared to the wearable-based solution. However, the existing approaches either require an ultra-wideband radio which is not commercially available or do not perform well in practical working environments. The millimeter-wave (mmWave) radio is a promising solution for fine-grained heartbeat and respiration sensing applications because of its directionality and sensitivity. However, we find traditional mmWave algorithms suffer from background noise in practical scenes. In this article, we address this issue by designing a robust algorithm for heart rate detection based on time-domain and frequency-domain information. We implement a phase-modulated system on the software-defined radio platform and evaluate the algorithm performance. Also, we evaluate the impact of several practical factors, such as detecting distance, aiming point, depression angle, human orientation and beam width on the proposed heart rate algorithm. Finally, we explore the feasibility of mmWave on vital sign detection with strong background interference and in scenes of real life where the antennas are hanging on the ceiling. The results show that the mean estimation error of respiration and heartbeats are 0.487 Bpm and 2.386 bpm.