Sleep Scoring with a UHF RFID Tag by Near Field Coherent Sensing

Long-term sleep scoring is very important in clinical settings to monitor patients’ recovery and at homes for both children and adults. In a cost-effective manner, quality of sleep can often be assessed by the upper-body movement together with heartbeat and respiratory monitoring. Instead of the conventional polysomnogram (PSG) which is uncomfortable due to skin contact of sensors and electrodes, this paper presents radio frequency (RF) near-field coherent sensing (NCS) by a single passive RF identification (RFID) tag in the chest area without requiring skin touch, where heart rates, breath rhythm, and motion can be synchronously extracted. Motion classification is based on support vector machine (SVM) with semi-supervised learning. Sudden body jerk, tossing, and turning can be recognized correctly in 91.06% of the test cases. The heart rate detection is also improved after motion artifact correction.