HRV Estimation In Electrocardiosignal Interferences Based On Photoplethysmography Signals

Heart rate variability(HRV) is produced in the periodical change of heart beat intervals, which is one of the important indices for reflecting the sympathetic nerve and vagus nerve activity’s balance in the autonomic nervous system. It can be used for many diseases’ prediction or diagnosis, however, HRV is derived from ECG signals,ECG signals’ acquisition needs many electrodes and multifarious attachment. At the same time, ECG signals acquisitioned by monitoring equipment often contain interferences caused by many factors, including human movement. Meanwhile, HRV in interferences is difficult to be directly extracted, which makes some disease detection methods or system based on HRV to be less reliable or even invalid. So, in order to improve the accuracy and real-timeliness of heart rate variability(HRV)estimation in electrocardiogram(ECG) signals interferences, a novel heart rate variability estimation method based on photoplethysmography(PPG) signals is proposed. The short-time autocorrelation principle is used to detect interferences in ECG signals, then, the fast power spectrum estimation method and the improving sliding window iterative Discrete Fourier Transform(DFT) are used to estimate HRV in ECG interferences from the synchronously acquisitioned PPG signals. Finally, the proposed methods of HRV estimation algorithm are compared with recently existing representative algorithms, respectively. The results show that the proposed methods in the paper are more accurate and more realtime.The following several aspects of the research work have mainly been completed in the paper:(1) The filtering of electrocardiogram(ECG) signals and photoplethysmography(PPG) signals. Due to ECG signals and PPG signals are both weak signals, usually introduce baseline drift, power frequency interference, myoelectricity interference,other noise and interferences in the process of acquisition. In order to filter out the interferences, at first, the normal ECG and PPG signals waveform are introduced, then,the integral coefficients filter is designed to filter these interferences, which can filter interferences more real-timely than IIR and FIR filters.(2) The detection of ECG signals interferences. According to the characteristics of the interferences in ECG signals, the principle of short-time autocorrelation is used to detect interferences rapidly. Through combining two dynamic parameters of the smooth degree and dynamic coefficient of variation, the interferences detectionaccuracy is improved. The international commonly used MIT-BIT Arrhythmia Database/Challenge 2014 Training Set(Challenge/2014/Set-p) data is used to verify the interferences detection algorithm proposed in the paper. In addition, the proposed interferences detection algorithm has been compared with the existing commonly used several kinds of interferences detection algorithm to assess the performance of the proposed algorithm. The results show that the proposed interferences detection algorithm is more accurate and more realtime.(3) The heart rate variability estimation of interferences in ECG signals. Using the correlation between ECG signals and PPG signals, the fast power spectrum estimation method and the improving sliding window iterative Discrete Fourier Transform(DFT) are used to estimate HRV in ECG interferences from the synchronously acquisitioned PPG signals, respectively. Then, the validity and reliability of the algorithms proposed are evaluated with the means of comparing with the current commonly used heart rate variability estimation algorithm.(4) The identification method of coronary artery disease based on HRV estimation. Taking the real-timeliness and accuracy of algorithm into account, the proposed improved sliding window iterative DFT algorithm is used to extract the HRV of healthy person and patients with coronary artery disease, respectively. According to the difference between them, their respective features in time domain and frequency domain are both extracted. In order to reduce the complexity of the algorithm, the best features combination is selected by t-test. Then the optimized support vector machine(SVM) is applied to classify them. Finally, the classification algorithm is compared with the commonly used several kinds of intelligent algorithm to analyze the advantages and disadvantages of various algorithms and evaluate the property of classification algorithm further.