A Novel Detection Method for Heart Rate Variability and Sleep Posture Based on a Flexible Sleep Monitoring Belt

• Published in: IEEE sensors journal Vol. 25; no. 3; pp. 5178 - 5191
• Main Authors: He, Chunhua, Liu, Shuibin, Fang, Zewen, Wu, Heng, Liang, Maojin, Deng, Songqing, Lin, Juze
• Format: Journal Article
• Language: English
• Published: New York IEEE 01.02.2025; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Accuracy; Angular velocity; Artificial neural networks; Autonomic nervous system; Ballistocardiography; Bioelectricity; Convolutional neural network (CNN) model; Edge computing; Electrocardiography; Electromechanical sensors; flexible sleep monitoring belt (FSMB); Heart rate; heart rate variability (HRV) detection; Inertial platforms; Inertial sensing devices; MEMS inertial measurement unit (IMU); Monitoring; Optical variables measurement; Pressure sensors; Seismic measurements; Sensor arrays; Sensors; Sleep; Sleep apnea; sleep posture detection; Telemedicine; Three-dimensional displays; Vibration measurement; Wavelet analysis; Wavelet packets; Wavelet transforms
• ISSN: 1530-437X; 1558-1748
• DOI: 10.1109/JSEN.2024.3518082

Heart rate variability (HRV) is an important indicator for assessing the function of the cardiac autonomic nervous system (ANS), and it is important for early detection and prevention of cardiovascular diseases, stress management, and mental health. Besides, different sleep postures have different effects on respiration and ventilation, and inappropriate sleep postures may lead to organ compression and obstructive sleep apnea (OSA). Therefore, HRV and sleep posture detection are very significant. However, there is a lack of the high-comfortable, low-cost, and high-accuracy detection methods. In this article, a novel detection method for HRV and sleep posture based on a flexible sleep monitoring belt (FSMB) is proposed. The test platform, including an FSMB and a bioelectrical signal acquisition circuit (BSAC), as well as the test flow, is described in detail. The BSAC composed of a series of amplifiers and filters is designed to acquire the electrocardiography (ECG) signal, while the FSMB mainly composed of a MEMS inertial measurement unit (IMU) and a pressure sensor array is designed to acquire the ballistocardiography (BCG) or gyrocardiography (GCG) signal. Besides, the HRV features of ECG, BCG, and GCG signals are extracted by the wavelet packet transform (WPT) analysis, and the short-time energies of the triaxial accelerations and angular velocities are extracted as the features for sleep posture detection. For facilitating the realization with edge computing, a lightweight convolutional neural network (CNN) model is proposed to recognize the sleep posture. The experimental results indicate that the detection accuracy of HRV with BCG signal is slightly bigger than that with GCG signal, reaching 91.1% compared with the result of ECG signal. In addition, the detection accuracy of sleep posture with the proposed CNN model achieves 96.44%. Therefore, the proposed detection method of HRV and sleep posture based on the FSMB is effective and feasible.