基于脉搏信号的心电干扰段心率变异性估计

发布时间：2018-01-15 17:35

本文关键词：基于脉搏信号的心电干扰段心率变异性估计　出处：《兰州理工大学》2016年硕士论文　论文类型：学位论文

【摘要】：心率变异性(Heart Rate Variability,HRV)产生于心脏搏动间期的周期性变化,是反映自主神经系统交感神经与迷走神经活性及其平衡性的重要指标,可用于许多心血管疾病和非心血管疾病的预测和诊断。通常,HRV从心电信号中获取,但是获取心电信号的电极导联和连线繁杂,同时,心电信号监护设备所采集的信号经常含有人体运动等因素引起的干扰段,干扰段的HRV难以直接提取,从而导致一些以HRV为基础的疾病检测方法或系统产生误判,可靠性降低。因此,为了提高心电信号干扰段检测的准确性和实时性,且准确估计干扰段信号的HRV,本文提出了一种新的心电干扰段HRV估计方法。该方法采用短时自相关原理检测心电信号中的干扰段,并利用脉搏信号与心电信号间的相关性,采用基于自相关函数的快速功率谱估计法和改进的滑窗迭代DFT算法,分别从同步采集的脉搏信号中估计心电干扰段的HRV,并与现有的典型算法进行比较。结果表明,文中提出的HRV估计算法具有较高的准确性和较强的实时性。本文主要研究了以下四个方面:(1)心电脉搏信号的滤波。心电脉搏信号都为微弱的信号,在它们的采集的过程中会引入各类噪声和干扰。通常采集的心电脉搏信号中都会含有基漂、肌电干扰和50Hz的工频干扰。为了滤除这些干扰和噪声信号,本文首先对标准的心电脉搏信号波形进行了介绍,然后设计实时性比IIR和FIR滤波器更强的整系数滤波器进行滤波。(2)心电信号干扰段检测。根据心电信号干扰段的特点,采用短时自相关原理快速检测心电信号的干扰段,并结合光滑度和动态变异系数两个动态参数提高干扰段检测的准确度。采用国际常用MIT-BIT Arrhythmia Database/Challenge2014 Training Set(challenge/2014/set-p)数据,对提出的干扰段检测算法和现有常用的几种干扰段检测算法进行仿真比较,评估所提算法的性能。结果表明文中提出的干扰段检测算法的准确度更高并且实时性更强。(3)心电干扰段的HRV估计。利用心电信号和脉搏信号之间的相关性,采用基于自相关函数的快速功率谱估计法和改进的滑窗迭代DFT算法,分别从同步采集的脉搏信号中估计心电干扰段的HRV。并将文中提出的HRV估计算法与目前常用的HRV估计算法进行仿真比较,评估文中算法的有效性和可靠性。(4)基于心率变异性估计的冠心病识别方法研究。考虑到算法的实时性和准确性,采用文中提出的改进滑窗迭代DFT算法提取健康人和冠心病患者的HRV,针对它们的不同,分别提取它们时域和频域特征,然后进行特征选择,选取最佳的特征组合,采用优化的支持向量机对选取的特征组合分类。最后,与常用的几种智能算法进行实验比较,分析各种算法的优缺点,对文中的分类算法进行进一步性能评估。
[Abstract]:Heart rate variability (HRV) is produced by periodic changes in heart beat interval. It is an important indicator of sympathetic and vagal nerve activity and its balance in autonomic nervous system. It can be used in the prediction and diagnosis of many cardiovascular diseases and non-cardiovascular diseases. At the same time, the ECG signal collected by ECG monitoring equipment often contains interference segment caused by human movement and other factors, the HRV of interference segment is difficult to be directly extracted. Therefore, in order to improve the accuracy and real time of ECG interference detection, some disease detection methods or systems based on HRV cause misjudgment and lower reliability. In this paper, a new HRV estimation method for ECG interference segment is proposed, which uses the principle of short time autocorrelation to detect the interference segment in ECG signal. Using the correlation between pulse signal and ECG signal, the fast power spectrum estimation method based on autocorrelation function and the improved sliding window iterative DFT algorithm are adopted. The HRV of ECG interference segment is estimated from synchronously collected pulse signals, and compared with the existing typical algorithms. The results show that. The HRV estimation algorithm proposed in this paper has higher accuracy and better real-time. This paper mainly studies the filtering of ECG pulse signal in the following four aspects: 1) ECG pulse signal are all weak signals. All kinds of noise and interference will be introduced in the process of their acquisition. The ECG pulse signals usually collected will contain fundamental drift, electromyography interference and 50 Hz power frequency interference. In order to filter out these interference and noise signals. In this paper, the standard waveform of ECG pulse signal is introduced. Then, an integer coefficient filter, which is stronger than IIR and FIR filters in real time, is designed to detect the interference segment of ECG signal. According to the characteristics of the interference segment of ECG signal. The short time autocorrelation principle is used to detect the interference segment of ECG signal. The accuracy of interference detection was improved by combining the two dynamic parameters of smoothness and dynamic coefficient of variation. MIT-BIT Arrhythmia was adopted in this paper. Database/Challenge2014 Training Settings (. Challenging / 2014 / set-p data. The proposed jamming segment detection algorithm and several commonly used jamming segment detection algorithms are simulated and compared. The performance of the proposed algorithm is evaluated. The results show that the proposed interference segment detection algorithm is more accurate and real-time. HRV estimation of ECG interference segment. Correlation between ECG and pulse signals was used. The fast power spectrum estimation method based on autocorrelation function and the improved sliding window iterative DFT algorithm are adopted. HRVs of ECG interference segment are estimated from synchronously collected pulse signals, and the HRV estimation algorithm proposed in this paper is simulated and compared with the HRV estimation algorithm commonly used at present. Evaluation of the effectiveness and reliability of the algorithm in this paper. (4) Research on the recognition method of coronary heart disease based on heart rate variability estimation, taking into account the real-time and accuracy of the algorithm. The improved sliding window iterative DFT algorithm proposed in this paper is used to extract the HRVs of healthy persons and patients with coronary heart disease (CHD). According to their differences, the time-domain and frequency-domain features are extracted, and then the feature selection is carried out. The best feature combination is selected, and the selected feature combination is classified by the optimized support vector machine. Finally, the advantages and disadvantages of each algorithm are analyzed and compared with several common intelligent algorithms. Further performance evaluation of the classification algorithm is carried out.
【学位授予单位】：兰州理工大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：R540.4

【相似文献】