Non-invasive blood pressure measurement algorithm using neural networks

Abstract

The oscillometric method is the most commonly used automatic monitoring blood pressure measurement method nowadays. Height-based and Slope-based criteria are the two general means used to determine the systolic and diastolic pressures; however they are disputed for their accuracy. Thus, the auscultatory method continues to be the gold-standard for these measurements. In this paper a newly developed cuff with piezofilm sensors and a pressure sensor to collect signals from the brachial artery is investigated. Using Neural Networks to classify the acquired pressure signals in various regions, an algorithm is developed and implemented in signal processing and heart beat/heart rate detection software. The algorithm is tested on 258 measurements from 86 subjects and shows good conformance to the standards set out by the Association for the Advancement of Medical Instrumentation and British Hypertension Society grade A criteria.