基于压电感知的科学睡眠监护系统

发布时间：2018-03-19 20:48

本文选题：睡眠监护　切入点：睡眠分期　出处：《电子科技大学》2016年硕士论文　论文类型：学位论文

【摘要】：失眠或者睡眠质量不高已经成为影响现代人健康的一大问题。获取并分析睡眠分期是科学评价睡眠质量、寻找影响睡眠因素和改善睡眠状况基础。传统的自动睡眠分期方法大多是针对脑电信号,然而脑电信号的采集需要借助精密的仪器和复杂的电极,无法满足家庭睡眠监护的需求。压电感知床垫能够将睡眠监测对睡眠的影响降到最低,是非常适合家用睡眠监护的解决方案,但是床垫采集的BCG信号是心跳、呼吸和体动的混合信号,体动信号会对提取心跳和呼吸信号的特征产生不可忽视的影响,而体动信号本身也对睡眠分期有显著的价值。大多数的非脑电自动睡眠分期算法是基于其中某一种单一信号来分期的,很多方法采用的是通过多导睡眠图收集的标准的生理电信号,而对长时间的睡眠压电混合信号进行分期不一定有较高的准确率。本文提出一种融合多种生理数据的睡眠分期算法,基于床垫采集的压电数据进行睡眠分期,同时实现了基于压电感知的移动睡眠监护系统,为实现家庭睡眠监测和远程睡眠监测提供了一套科学的解决方案。本文提出的融合多种生理数据的睡眠分期算法,首次将隐马尔可夫模型运用到呼吸信号的睡眠分期应用当中,证明隐马尔可夫模型可以识别呼吸在不同睡眠时相中存在不同的模式和规律。本文提出的算法通过分析心率和呼吸率的隐马尔可夫模型在识别不同睡眠分期时结果的优势与不足,综合两者的优点改进原算法,同时设计了体动检测算法与之结合,校正因为体动影响到心率、呼吸率特征而出现的错误分期,最终实现了对觉醒期、浅睡眠期、深睡眠期和REM期的自动识别以及体动的记录,达到准确率66.91%,并在与同类产品的比较中有较好的表现。本文搭建了基于压电感知的移动睡眠监护软件平台,为实现家庭睡眠监测和远程睡眠监测提供了一套科学的解决方案。软件系统以基于Android平台和Web技术,定义了睡眠生理数据的通信方案和与用户的交互方式,集成了融合多种生理数据的睡眠分期算法和BCG信号的预处理方法,实现了远程睡眠监测、睡眠结构可视化、睡眠质量分析、睡眠交流和睡眠知识等个性化功能。
[Abstract]:Insomnia or poor sleep quality has become a major problem affecting the health of modern people. Obtaining and analyzing sleep stages is a scientific evaluation of sleep quality. Traditional automatic sleep staging methods are mostly aimed at EEG signals, but EEG acquisition requires sophisticated instruments and complex electrodes. It can't meet the needs of family sleep monitoring. Piezoelectric sensing mattresses can minimize the impact of sleep monitoring on sleep, which is a perfect solution for home sleep monitoring, but the BCG signals collected by the mattresses are heartbeats. A mixture of breathing and body movement, which can have an unnegligible effect on extracting the characteristics of heartbeat and respiratory signals. The body motion signal itself has significant value for sleep staging. Most non-EEG automatic sleep staging algorithms are based on one of the single signals. Many methods use standard physiological signals collected through polysomnography. In this paper, a sleep staging algorithm is proposed, which is based on the piezoelectric data collected from the mattress. At the same time, a mobile sleep monitoring system based on piezoelectric perception is implemented, which provides a scientific solution for home sleep monitoring and remote sleep monitoring. For the first time, the hidden Markov model is applied to the sleep stages of respiratory signals. It is proved that the hidden Markov model can recognize the different patterns and rules of breathing in different sleep phases. The algorithm proposed in this paper is based on the analysis of heart rate and breathing rate of hidden Markov models to identify different sleep stages. The advantages and disadvantages of fruit, Combining the advantages of the two methods, we have improved the original algorithm, designed the body motion detection algorithm and combined it with it, corrected the wrong stages caused by the influence of body movement on heart rate and breathing rate, and finally realized the phase of awakening and shallow sleep. The automatic recognition of deep sleep period and REM phase and the recording of body movement can achieve the accuracy of 66.91and have good performance in comparison with the similar products. In this paper, a mobile sleep monitoring software platform based on piezoelectric perception is built. It provides a scientific solution for home sleep monitoring and remote sleep monitoring. Based on Android platform and Web technology, the software system defines the communication scheme of sleep physiological data and the way of interaction with users. A series of sleep staging algorithms and BCG signal preprocessing methods are integrated to realize the personalized functions of remote sleep monitoring, sleep structure visualization, sleep quality analysis, sleep communication and sleep knowledge.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：R740;TN911.6

【参考文献】