基于指端脉搏波的血压测量方法及其在智能终端的初步实践

发布时间：2018-06-08 04:15

本文选题：指端脉搏波 + 血压测量　；参考：《浙江大学》2017年硕士论文

【摘要】：基于脉搏波测量血压的方法是现阶段无创连续血压测量的重要研究分支。基于脉搏波的血压测量方法包括基于脉搏波传导速度的测量方法、基于脉搏波传导时间的测量方法和基于脉搏波特征参数的测量方法。前两种测量方法需要采集两路人体生理信号,而两路信号间的稳定性对测量结果有较大的影响。后一种通过分析脉搏波特征参数与血压之间的关系,建立脉搏波特征参数-血压计算模型,实现对每搏血压的连续测量。现阶段基于脉搏波特征参数的血压测量方法主要集中于对肱动脉和桡动脉脉搏波的研究,血压计算模型多是通过回归分析方法建立的线性回归方程,模型一般适用于特定的测量个体和脉搏波采集设备。指端脉搏波与胲动脉、桡动脉脉搏波具有较好的相似性,且更容易被测量,已被用于各类心血管生理参数的计算。移动智能终端软硬件性能的快速提升,使得基于指端脉搏波进行血压测量的便捷性得到了极大提高,具有良好的应用前景。本文研究了指端脉搏波特征参数与血压之间的关系,提出了一种基于指端脉搏波的血压测量方法,并在移动终端上进行了初步实践。首先,基于对指端脉搏波和现有的脉搏波特征参数的分析,提出了本文采用的指端脉搏波特征参数,并通过分析各参数与血压之间的相关性,为利用这些参数计算血压提供了理论依据。然后,本文给出了针对两种脉搏波的获取、预处理和特征参数提取方法。一种脉搏波取自MIMIC数据库,用于血压计算模型的建立和评估;另一种脉搏波利用智能手机采集,用于检验血压计算模型在实际场景下的性能。其次,本文给出了血压计算模型的建立和评估方法。本文采用偏最小二乘回归分析和BP神经网络两种方法建立脉搏波特征参数-血压计算模型,前者能在自变量间存在较高相关性的情况下拟合自变量和因变量间的关系;后者则能够很好地拟合自变量和因变量间的非线性关系。评估结果显示,利用BP神经网络建立的模型具有更好的血压预测能力。最后,基于当前移动端慢病管理的背景,将本文建立的模型在移动端进行了初步实践——开发了一个基于Android的血压测量模块,并在已有的移动端慢病管理平台进行了初步应用。
[Abstract]:Pulse wave based blood pressure measurement is an important branch of noninvasive continuous blood pressure measurement. Pulse wave based blood pressure measurement method includes pulse wave velocity measurement method, pulse wave conduction time measurement method and pulse wave characteristic parameter measurement method. The first two methods need to collect two human physiological signals, and the stability between the two signals has a great influence on the measurement results. By analyzing the relationship between pulse wave characteristic parameters and blood pressure, a pulse wave characteristic parameter-blood pressure calculation model is established to realize the continuous measurement of stroke blood pressure. At present, blood pressure measurement methods based on characteristic parameters of pulse wave mainly focus on the study of brachial artery and radial artery pulse wave. The calculation models of blood pressure are mostly linear regression equations established by regression analysis. The model is generally suitable for specific measuring individuals and pulse wave acquisition equipment. The fingertip pulse wave has good similarity with hydroxylamine artery and radial artery pulse wave and is more easily measured. It has been used to calculate various cardiovascular physiological parameters. With the rapid improvement of hardware and software performance of mobile intelligent terminal, the convenience of blood pressure measurement based on fingertip pulse wave is greatly improved, and it has a good application prospect. In this paper, the relationship between the characteristic parameters of fingertip pulse wave and blood pressure is studied, and a blood pressure measurement method based on finger pulse wave is proposed, and the preliminary practice is carried out on mobile terminal. Firstly, based on the analysis of the characteristic parameters of finger pulse wave and existing pulse wave, the characteristic parameters of finger pulse wave are put forward, and the correlation between these parameters and blood pressure is analyzed. It provides a theoretical basis for the calculation of blood pressure using these parameters. Then, this paper presents two pulse wave acquisition, preprocessing and feature extraction methods. One pulse wave is taken from the miMIC database for the establishment and evaluation of the blood pressure calculation model, and the other pulse wave is collected by smart phone to test the performance of the blood pressure calculation model in the actual scenario. Secondly, the establishment and evaluation method of blood pressure calculation model are given. In this paper, two methods of partial least square regression analysis and BP neural network are used to establish a pulse wave characteristic parameter-blood pressure calculation model. The former can fit the relationship between independent variables and dependent variables when there is a high correlation between independent variables. The latter can fit the nonlinear relationship between independent variables and dependent variables well. The evaluation results show that the BP neural network model has better blood pressure prediction ability. Finally, based on the background of the current mobile side slow disease management, the model is applied in the mobile side. A blood pressure measurement module based on Android is developed and applied to the existing mobile side chronic disease management platform.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP183;R443.5

【参考文献】