Research Of Central Arterial Pressure Estimating Method Based On Pulse Wave

The biological principle of central arterial pressure is the lateral pressure on the blood vessels at the root of the ascending aorta.Central arterial pressure is more closely related to cardiovascular mechanism and physiology than peripheral arterial pressure.Central arterial pressure can more direct and More accurate responses to coronary,left ventricular and cardiovascular pressure states.Therefore,central arterial pressure has always been one of the important medical problems that people pay close attention to and study.There is a strong correlation between pressure and pulse wave.Two problems need to be solved to correctly infer the central arterial pressure: First,the analysis and identification of tidal wave is the premise of these calculations and derivation.But the identification of obvious tidal wave of pulse wave has been perfected,and the research on the identification of hidden tidal wave is relatively few.Second,the design and training of general transfer function,Universal transfer function(GTF)is the key to calculate the central arterial pressure waveform.In this paper,the problem of the discrimination of covert tidal waves in pulse waves is studied.The rich and varied part of pulse waves is hidden in smooth waves,and its peak is not obvious.The accuracy of the traditional method is low.This paper begins with the analysis of the frequency component of the pulse wave and finds the distribution range of the tidal wave in frequency domain.Then the pulse wave is decomposed by sym4 wavelet basis.The pulse wave is decomposed into several frequency components by haar wavelet basis,and the tidal wave is identified by subtracting the wavelet coefficients of the two different wavelet basis functions.In the process,the recognition effect of different layers is studied through experiments.Finally,the difference between the third layer decomposition coefficients of sym4 and haar is identified with the highest accuracy.The third peak of the third order difference wave from the third to the tenth peak point corresponding to the pulse wave position of the sixth over zero is a concealed tidal wave position.Through the MIMIC intensive care data(MIT Medical School of the Massachusetts Institute of Technology)and self-collected data validation,This method has obvious effect on the recognition of obvious tidal wave and covert tidal wave.Through the verification of database data and self-mining data,the recognition accuracy of obvious tidal wave is 97.5 perscent and that of covert tidal wave is 91.11 perscent.It is also necessary to get the help of the central arterial pressure wave to estimate the central arterial pressure by the corresponding blood pressure at the tidal wave.At present,it is difficult to obtain the training data for estimating the transfer function of the central artery waveform from the radial artery waveform.In this paper,the data of pulse wave and pressure wave of central artery are used in MIT MIMIC.After comparing and analyzing the results of Z transform,Laplace transform and Fourier transform,the general transfer function(GTF)is trained by Fourier transform.The converted central arterial pressure waveform is accurate incalculating the central arterial pressure,and meets the requirements of positive and negative 5mmHg errors.