Remote Respiratory Variables Tracking With Biomedical Radar-Based IoT System During Sleep
Shuqin Dong, Wen Li, Yuchen Li, Jingyun Lu, Zhi Zhang, Chengmei Yuan, Changzhan Gu
Abstract
Nocturnal respiratory function assessment is essential for inpatients’ daily healthcare. However, the nasal cannula commonly used in hospitals can affect the patient’s breathing pattern and cause discomfort to the patient. A radar-based contactless detection technique provides a noncontact monitoring way to overcome such limitation. However, to be an alternative to airflow, the radar system requires highly accurate respiratory waveform detection. In this article, a custom-designed biomedical radar-based Internet of Things (IoT) system integrated with automatic respiratory variables identification algorithm is developed to provide accurate long-term measurement of respiratory phases and amplitude during sleep. The overnight experimental data from biomedical radar and polysomnography was simultaneously recorded for ten adults, a total duration of approximately 3851 min. Data from adults with suspected sleep apnea were also included. Considering the instability of respiratory activity during REM sleep, we compared the performance of the system’s measurements during REM and non-REM sleep. The results demonstrate the continuous measurement can provide the long-term respiratory variables for sleep analysis. The total experimental measurement reveals a remarkable 97%, 93%, and 92% accuracy in respiration-to-respiration interval (RRI), inhale duration (ID), and exhale duration (ED), respectively, which provides accurate and convenient monitoring method and might be a part of medical IoT in the future.