| Title |
A Real-time Blood Pressure Estimation System for Smart Healthcare on Wearable Device |
| Authors |
이정철(Jung Chul Lee) ; 이재학(Jeahack Lee) ; 김병수(Byung-Soo Kim) ; 전석훈(Seokhun Jeon) |
| DOI |
https://doi.org/10.5573/ieie.2022.59.12.111 |
| Keywords |
Embedded; Smart healthcare; Light-weight AI; Blood pressure estimation; Wearable |
| Abstract |
The importance of accurate hypertension diagnosis has been emphasized due to the rapid aging of the population and increasing cardiovascular diseases caused by hypertension. Although machine learning and deep learning-based algorithms are being studied for fast and accurate blood pressure estimation, it is difficult to implement in embedded systems with hardware limitations such as power consumption and computational complexity. In this paper, we propose a piecewise linear blood pressure estimation algorithm with high accuracy and implementing this algorithm considering task scheduling on a watch-type wearable platform for real-time blood pressure estimation. The proposed system has verified the real-time blood pressure estimation operation and shows an average of about 32% improvement in the MIMIC standard blood pressure estimation error blood pressure(5, 10, 15 mmHg). In addition, the average measurement error and standard deviation with household blood pressure machine satisfying the medical device standard demonstrated similar performance. |