A Non-contact and Unconstrained Sleep Health Monitoring System

Abstract

Clinically, polysomnography (PSG) is used to assess sleep quality by monitoring various parameters, such as Electroencephalogram (EEG), electrocardiogram (ECG), Electrooculography (EOG), Electromyography (EMG), pulse, oxygen saturation, and respiratory rate. However, in order to assess these parameters, PSG requires a variety of sensors that must make direct contact with patients’ bodies, which can affect patients’ quality of sleep during testing. Thus, the use of PSG to assess sleep quality can yield invalid and inaccurate results. To address this gap, this paper proposes a sleep health monitoring system that has no restraints and does not interfere with sleep. This method collects ballistocardiogram (BCG) due to ejection by placing a piezoelectric film sensor under a sleeping cushion and evaluates three indicators: heart rate variability (HRV), respiration, and body movements. The ECG and BCG of 10 subjects were collected synchronously while the subjects were lying flat. Specifically, power-line interference was eliminated by adaptive digital filtering with a minimum mean square. A paired t-test revealed that there were no significant differences between BCG and standard ECG signals in the time-domain, frequency-domain, and nonlinear parameters of HRV. Respiration and body motion were extracted from the BCG in order to effectively monitoring of sleep apnea and nighttime bed-off times. Compared with other non-contact monitoring methods, such as acceleration sensors, coupling electrodes, Doppler radar and camera, the system presented in this paper is superior, as it has high signal quality, strong anti-interference ability, and low cost. Moreover, it does not interfere with normal sleep.