Abstract：The effective monitoring of sleep posture provides a direct reference for diagnosis, prevention
and treatment of some disease. This paper proposed a new method based on bioelectrical impedance
technology for sleep posture recognition. To recognize sleeping posture, this method collected
information of sleeping posture from the respiratory impedance signal using non-invasive, continuous
bioelectrical impedance technology around the chest. The hardware part acquired human’s respiratory
signal through two channels multi-channel switch around the chest in real-time based on Agilent E4980A
electrical impedance instrument. The software part was composed of Labview which acted as a realtime
display and storage online system and the Matlab which could classify and recognize the algorithm
offline. In this study, 19 volunteers were recruited to participate experiment for 4 cycles. All of the
volunteers were separately perform normal breathing exercises in posture of supine position, left lateral
position, right lateral position and prone position. Support vector machine (SVM) was used to recognize
and classify these 4 kinds of postures according to the extraction of characteristic values. The results
showed that the highest classification accuracy of test group could reach 94.6% by using SVM.
许欢，张平. 基于生物电阻抗技术的睡眠姿势识别方法的探讨[J]. 中国医疗设备, 2017, 32(6): 39-44.
XU Huan, ZHANG Ping. Study on Sleep Posture Recognition Based on Bioelectrical Impedance
Technology. China Medical Devices, 2017, 32(6): 39-44.
