A New Blood Pressure Estimation Method Based on Neural Network Algorithm Model

2018 International Conference on Advanced Materials, Intelligent Manufacturing and Automation

材料科学;机械制造;运输工程

Lu, Qingqiang^1 ; Dai, Erhan^1

College of Automation, Nanjing University of Posts and Telecommunications, Nanjing

210023, China^1

关键词: Blood pressure estimation;    Feature points extraction;    Hypertensive patients;    Measurement accuracy;    Neural network algorithm;    Pressure calculation;    Pulse wave transit time;    Signal preprocessing;   

Others  :  https://iopscience.iop.org/article/10.1088/1757-899X/382/5/052027/pdf
DOI  :  10.1088/1757-899X/382/5/052027

学科分类：材料科学（综合）

来源: IOP

【 摘 要 】

Measuring blood pressure is one of the main approach of assessing cardiovascular status and heart disease. The method of using PWTT to estimate blood pressure in clinical practice has matured. However, some uncontrollable subtle changes are likely to affect the measurement accuracy of pulse wave transit time. To solve the above problems, a new blood pressure estimation method based on neural network algorithm model is proposed based on the data of a hospital in Nanjing. The blood pressure is first calculated in advance by traditional PWTT algorithm through data analysis, and then modified with the algorithm of BP neural network algorithm. The method is divided into four steps: signal preprocessing and feature points extraction, PWTT computing BP value, neural network algorithm, 4 significant factors selection and blood pressure calculation, and finally combined with PWTT algorithm to estimate the final blood pressure. Compared with the real blood pressure data of 100 hypertensive patients, the error of our new method is less than 5 mmHg and the standard deviation is less than 9 mmHg. The results show that using this method to estimate blood pressure can meet the clinical needs and has good practical value.