Photoplethysmogram-based heart rate and blood pressure estimation with hypertension classification

Abstract

A photoplethysmogram (PPG) is an optically-derived signal that records the variation in blood volume within the microvasculature. Certain cardiovascular diseases (CVDs) are symptomatic of damaged blood vessels and problems in blood flow, including hypertension. While software implementations for heart rate and blood pressure estimation exist, point-of-care systems demand hardware-based implementations for real-time estimations to be useful for CVD detection. In this study, digital field programmable gate array (FPGA) based systems are developed for heart rate and blood pressure estimation from PPG signals by means of linear regression. In addition to the blood pressure estimation system, we present a prototype hypertension level detection system that achieves 92.42% accuracy while consuming 0.364 W of power. The Mean Absolute Error (MAE) $\pm$ Standard Deviation (SD) for heart rate estimation is 3.17 ± 2.79 beat per minute. The corresponding results for systolic and diastolic blood-pressure estimation are 4.75 ± 2.78 and 3.34 ± 2.60, respectively. The prototype can be further extended to wearable devices and medical equipment in the future.