A Respiration-Derived Posture Method Based on Dual-Channel Respiration Impedance Signals

• Published in: IEEE access Vol. 5; pp. 17514 - 17524
• Main Authors: Liu, Guanzheng, Li, Kunyang, Zheng, Lianrong, Chen, Wen-Hui, Zhou, Guangmin, Jiang, Qing
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.01.2017; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: backward elimination (BE); Impedance; Instruments; Lungs; Monitoring; Posture; Quality assessment; Radial basis function; radial basis function (RBF); Respiration; Respiration derived posture (RDP); Respiratory impedance; Sensors; Sleep; Sleep apnea; support vector machine (SVM); Support vector machines
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2017.2737461

Sleep posture has been used as a sleep assessment indicator in home health monitoring and clinical monitoring in recent years. Considering comfort and usability, unobtrusive sleep posture detection is needed. In this paper, we proposed a novel sleep respiration-derived posture (RDP) method based on left and right lung respiration impedance signals. We developed a dual-channel respiratory impedance acquisition system with wireless transmission. Then, support vector machine using radial basis function kernel was applied to recognize four typical sleep postures. Moreover, the performance of the SVM classifier was improved by using backward elimination. In-situ experiments with 16 subjects indicated that the RDP method reached an accuracy of 99.67%. Thus, our method is reliable in sleep posture recognition. Furthermore, a whole-night monitoring system based on respiration impedance can be conducted for sleep quality assessment.