A robust and flexible pulse wave sensory array enabling real-time non-invasive blood pressure monitoring

Abstract

Chronical cardiovascular decreases such as hypertension requires real-time and continuous monitoring of blood pressures (BPs). Pulse wave that contains critical and ample information on cardiovascular dynamics is a direct vital sign to extract blood pressure and therefore an epidermal wearable device enabling real-time acquisition of pulse waves becomes necessary. In this work, we propose and study a flexible pulse wave sensory array aiming for real-time wearable pulse wave acquisition with robustness. A piezoelectric sensor together with a thin-film transistor-based sensor interface circuit is used to detect multiple pulse waveforms at the location of radial artery, which are input to a convolutional neural network (CNN) for a deep-learning BP estimation model training. The estimated BPs are assessed by a correlation study with the BPs measured by conventional sphygmomanometers. Our study shows a reasonable accuracy of mean deviation (MD) and standard deviation (STD) of 2.84±7.53 mmHg for systolic BP (SBP) and 0.88±7.06 mmHg for diastolic BP (DBP).