Adaptive Continuous Measurement Of Blood Pressure Based On Pulse Condition Classification

Blood pressure is an important indicator of human health,especially the continuous blood pressure,it can indirectly reflect the operation status of the heart and blood vessels,which is an important criterion for clinical curative judgment of disease diagnosis,disease prevention.But now the continuous blood pressure measurement equipment in the market,mainly for wearable electronic sphygmomanometers,is lack of calculation accuracy,leading to human risk of morbidity by false judgment.The accurate continuous measurement of blood pressure and the effective judgment of the abnormal condition play important roles in the prevention of cardiovascular diseases and the prevention of hypertension.Therefore,in view of the problems above,a new method f-or adaptive continuous measurement of blood pressure based on pulse condition classification is put forward in this paper.This method uses the new sensor-RF radar to acquire the dual signals of radial artery of human,and then hierarchical associative model is introduced to realize the implementation of pulse condition classification.Finally,the blood pressure is measured in real time by a two-stages adaptive BP prediction model.The main contents of this paper include:(1)Understand the working principle of RF-radar and its advantages,design human radial artery pulse wave acquisition system and data display and storage system based on Labview.The effectiveness of the designed system is verified by comparing with the critical pulse wave acquisition system.(2)Accurate classification of pulse condition is the basis of blood pressure prediction,only accurate classification can ensure the blood pressure prediction.Set up the stereotype associative mechanism model based on the stereotype thinking mechanism of human,and then pulse condition classification model based on hierarchical stereotype associative model is introduced.Pulse condition classification needs hierarchical implementation,the first step is to achieve rough classification and reduce the number of guiding directions,and effective pulse condition classification is obtained by stereotype associative neural network.This model reflects the strong stereotype associative ability through the neuronal interaction association network,then auto associative process is done between the testing pulse and the typical pulses in memory base with the guided mutation and pulse evolution rules.(3)Two-stages adaptive BP prediction model is proposed in the estimation process of blood pressure.Establish the linear models of the blood pressure based on the internal relationships between pulse conditions and blood pressure.According to the pulse condition and the information of the person,the blood pressure value can be adjusted dynamically based on the first stage of blood pressure model.Then,the second stage blood pressure value can be adjusted and calculated by the PSO-BP neural network with parameter library.The experimental results show that the pulse condition classification accuracy of five common pulse conditions based on hierarchical stereotype model can reach 92.86%,higher than any other classification method.At the same time,the accuracy of the adaptive blood pressure prediction model can reach 94.65%,which can achieve the accurate judgment of abnormal blood pressure,even achieves great continuous tracking trends on individuals.