Noninvasive Continuous Blood Pressure Measurement Based On PPG

With the development of society,our country’s domestic population is becoming more and more aging,the quality of life and work pressure is increasing,and the high fat intake and low exercise intensity have led to the development of high blood pressure into the most dangerous cardiovascular system disease.In order to understand the individual’s blood pressure status better and reduce the threat of cardiovascular disease to health,it is necessary to continuously monitor blood pressure.It has become the focus of current research to achieve continuous dynamic blood pressure detection without affecting the normal physiological activities of the human body.Traditional blood pressure measurement schemes often have problems such as complicated operation and poor portability,which are not conducive to continuous real-time blood pressure measurement.The photoplethysmographic pulse wave signal uses light signals to obtain blood flow information of peripheral tissues,which can accurately reflect the changes of pulse wave.Collecting pulse wave signals in this way has the characteristics of simple operation,low cost and stable performance,and is suitable for real-time non-invasive continuous blood pressure measurement.This paper designs a feature point recognition algorithm based on PPG signal,and uses the correlation between pulse wave morphology characteristics and blood pressure,proposes an improved blood pressure detection model based on multiple linear regression,and designs and implements a non-invasive continuous blood pressure detection system.The main work and results of the thesis are as follows:1.Research on the fast extraction algorithm of feature points extraction algorithm based on PPG using incremental merge segmentation.This paper proposes a novel real-time algorithm for segmenting PPG into pulses and classifying artifacts.The incremental merge segmentation(IMS)algorithm is used to extract potential PPG morphological feature points,and then the PPG units are screened by using the ITA(Inverted Triangular Area)value in the PPG morphology,and these potential feature points are classified and identified in combination with an adaptive threshold method.The classified feature points can be used to estimate a variety of physiological indicators including blood pressure.The average detection sensitivity of the proposed algorithm for PPG units is over 85%,and the detection error rate for various feature points is controlled below 0.01%.2.Research on New Blood Pressure Calculation Model-Blood Pressure Detection Algorithm Based on Multiple Linear Regression-Correlation Weighted Selection(MLR-CWS).This paper proposes an improved method based on multiple linear regression.In the parameter selection of the model,it is screened on the basis of the correlation between each characteristic parameter and blood pressure,and divided into two types:basic characteristic type and supplementary characteristic type,respectively participating in the establishment of basic value calculation model and supplementary value calculation model.The blood pressure calculation scheme of basic value+deviation value is introduced,and accurate monitoring of blood pressure is realized through weighted quantization of blood pressure deviation value and correction of the basic value.Comparing the traditional multiple linear regression model and the blood pressure detection model based on PTT,the proposed algorithm model has obvious mean absolute error(MAE),standard deviation(STD),correlation strength and consistency test in back-judgment analysis.Advantage.3.Design and implementation of the non-invasive blood pressure continuous detection system.The system has the function of real-time display of signals and physiological indicators,and is compatible with multiple types of hardware devices.It can not only realize the detection of physiological indicators based on PPG signals online,but also perform offline processing and analysis of data in public data sets such as MIMIC.In this paper,In order to verify the effectiveness of the algorithm,the proposed algorithm is applied in combination with the blood pressure continuous detection system designed for data acquisition and experimental analysis.Through the public data set,different algorithm models are deeply compared and summarized,which further proves the excellent performance of the proposed scheme.