生物雷达监测睡眠呼吸暂停综合症的初步研究

发布时间：2018-02-26 03:03

本文关键词： 生物雷达特征分析睡眠呼吸暂停综合症模式识别　出处：《第四军医大学》2012年硕士论文　论文类型：学位论文

【摘要】：应用生物雷达技术，可以隔着衣物、被褥等物体对人体的心跳、呼吸等生理运动进行非接触监测，与传统的接触式监测相比，这一过程会减少电极、导线、传感器等对被监测目标引起的不适感，还可以克服接触式监测方法对大面积烧伤、恶性传染病等病人监测的局限性。因此生物雷达技术在临床上有很广泛的应用前景。睡眠呼吸暂停综合症是一种临床上常见的疾病，对人类的健康有较大的危害。多导睡眠图监测是临床诊断睡眠呼吸暂停综合症的金标准，在诊断的过程中需要长时间将很多导线连接到病人的体表，会给病人带来很大不便以及身体上的不适感。因此我们提出将生物雷达技术应用于正常的睡眠呼吸与睡眠呼吸暂停的鉴别，为了实现这一目标，，我们主要完成了以下工作： 1.生物雷达采集呼吸信号的可靠性研究建立生物雷达检测呼吸信号和绑带式检测呼吸信号的同步检测系统，将生物雷达采集的呼吸信号与绑带式压力传感器方法采集的呼吸信号进行相关性研究，验证了生物雷达监测呼吸信号的可靠性，为后续的进一步研究奠定基础。 2.生物雷达采集的呼吸信号去噪应用生物雷达对人体长时间进行呼吸监测的过程中，会有一些其它的信号对呼吸信号产生干扰，因此我们采用基于Kaiser窗的低通滤波器和基于等波纹逼近法的低通滤波器对呼吸信号进行滤波，提取较为纯净的呼吸信号，并对这两种方法进行了比较。 3.生物雷达采集的呼吸信号的特征分析对于生物雷达采集的正常呼吸信号，我们提取了每个呼吸周期的最大值和最小值，并计算固定时长的呼吸能量。通过计算最值的间期可以得到呼吸频率，通过对呼吸能量的计算可以反映出呼吸运动的强弱。 4.睡眠呼吸暂停综合症的分析鉴别根据睡眠呼吸暂停综合症的发病特征模拟睡眠呼吸暂停综合症，用生物雷达采集呼吸信号，采用模式识别的方法对睡眠呼吸暂停与正常睡眠呼吸进行鉴别。本课题的主要创新点： 1.采用等波纹逼近的方法设计低通滤波器，对生物雷达采集的呼吸信号进行滤波，相比其他方法的滤波器，实现相同的效果所需的滤波器阶数更低。 2.对生物雷达采集的正常呼吸信号和模拟睡眠呼吸暂停时的呼吸信号进行特征值提取，并应用短时平均幅度、短时方差、短时频谱中某一点的频率分量三个特征向量对两种呼吸状态进行区分。
[Abstract]:With the use of biological radar technology, non-contact monitoring of physiological movements such as heartbeat, breathing, and so on can be carried out through clothing, bedding and other objects. Compared with traditional contact monitoring, this process will reduce electrodes and conductors. The sensor can overcome the limitation of contact monitoring for patients with large area burn and malignant infectious disease, so it has a wide application prospect in clinic. Sleep apnea syndrome is a common clinical disease that is harmful to human health. Polysomnography monitoring is the gold standard for clinical diagnosis of sleep apnea syndrome. It takes a long time to connect a lot of wires to the patient's surface. It can cause great inconvenience and discomfort to patients. Therefore, we propose to apply the biological radar technology to the identification of normal sleep breathing and sleep apnea. In order to achieve this goal, we have mainly accomplished the following tasks:. 1. Research on the reliability of respiratory signal acquisition by biological radar. A synchronous detection system is established for detecting respiratory signals by biometric radar and bandage. The correlation between the respiratory signals collected by biometric radar and the respiratory signals collected by the method of bandage pressure sensor is studied. The reliability of monitoring respiratory signals by biological radar is verified, which lays a foundation for further research. 2.Respiratory signal denoising from biological radar. Some other signals will interfere with the respiratory signals during the long period of breathing monitoring by using the biometric radar. Therefore, we use low-pass filter based on Kaiser window and low-pass filter based on equal-ripple approximation to filter respiratory signal, extract purer respiratory signal, and compare the two methods. 3. Characteristic analysis of respiratory signals collected by biological radar. For the normal respiratory signals collected by the biological radar, we extracted the maximum and minimum values of each respiration cycle, and calculated the respiration energy of the fixed period. The strength of respiratory movement can be reflected by the calculation of respiratory energy. 4. Analysis and identification of sleep apnea syndrome. According to the characteristics of sleep apnea syndrome, sleep apnea syndrome was simulated. The respiratory signals were collected by biological radar, and the normal sleep apnea and sleep apnea were identified by pattern recognition. The main innovation points of this subject are as follows:. 1. Using the equal-ripple approximation method to design the low-pass filter and filter the respiratory signal collected by the biological radar. Compared with the filter of other methods, the order of the filter needed to achieve the same effect is lower. 2. To extract the characteristic values of the normal respiratory signals and the respiratory signals of simulated sleep apnea, and to apply the short time mean amplitude and short time variance. The frequency components of a point in a short time spectrum are distinguished by three eigenvectors.
【学位授予单位】：第四军医大学
【学位级别】：硕士
【学位授予年份】：2012
【分类号】：R766;R318.0

【参考文献】