Fine-Grained Respiration Monitoring During Overnight Sleep Using IR-UWB Radar

Siheng Li, Zhi Wang, Fusang Zhang, Beihong Jin

Abstract: Recently, vital sign and sleep monitoring using wireless signals has made great progress. However, overnight respiration monitoring remains a challenge due to human unconscious and uncontrollable movements during sleep. In the paper, we explore the potential of an IR-UWB radar and implement a fine-grained overnight respiration monitoring prototype. Particularly, we exploit the complementarity between amplitude and phase of the radar signal to eliminate blind spots, thus improving the detection rate of overnight respiration monitoring. Moreover, we propose a circle fitting based phase restoration algorithm to correct the respiration depth distortion, and further recognize four respiration patterns (i.e., apnea pattern, Tachypnea pattern, Kussmaul pattern and rapid change pattern of respiration rate), thus enabling fine-grained respiration monitoring during overnight sleep. The experimental results show that our prototype achieves high respiration detection rates and accurate respiration rates, outperforming the two existing approaches. In addition, our prototype has captured the apnea pattern many times in the real sleep scenarios.