Noninvasive Continuous Blood Pressure Estimation Algorithm Based on Features of Pulse Waves

Hypertension has become one of the main threats to human health. How to accurately and conveniently obtain the information of blood pressure (BP) continuously is the prerequisite for effective prevention and treatment of hypertension. In this paper, we develop a new algorithm to realize continuous noninvasive BP estimation. Based on the pulse and BP waves in Medical Information Mart for Intensive Care (MIMIC) database, we exploit all available features (78 time-frequency features) of pulse waves extracted from different domains to build BP estimation model based on Elman neural network. In addition, to increase the accuracy of estimation, we use the Mean Impact Value (MIV) algorithm to select the features having the significant influence on BP prediction for each independent subject from the 78 features, and the Particle Swarm Optimization (PSO) algorithm is used to optimize network structure. The method complies with the Association for the Advancement of Medical Instrumentation (AAMI) standard in the estimation of systolic blood pressure (SBP) and diastolic blood pressure (DBP) values. The results indicate that the proposed algorithm for the noninvasive estimation of the BP has certain significance for promoting continuous BP monitoring using mobile health-care gadgets.