Blood flow is a vital signal of life. When this flow slows down or becomes unstable, it can lead to cardiovascular diseases and shock. However, accurately measuring blood flow has traditionally required hospital equipment. Prof. Kyeongha Kwon’s research team at the School of Electrical Engineering has developed a wireless electronic patch that can measure blood flow in real time simply by attaching it to the skin.
The wireless wearable blood flow monitoring system developed by Prof. Kwon’s team combines deep learning (AI) with multilayer thermal sensing technology. This device can simultaneously measure blood flow velocity and blood vessel depth without directly contacting the vessels (a non-invasive method). Because sensor signals vary depending on how deeply the blood vessels are located beneath the skin, depth information is a key variable for accurately calculating blood flow.
Previously, ultrasound or optical methods were mainly used, but these approaches had limitations, as the equipment was large or the accuracy decreased depending on vessel depth. To overcome these limitations, the research team focused on the fact that when blood flows, subtle heat transfer occurs in the surrounding tissue.
The team developed a “multilayer thermal sensing” technology that analyzes heat transfer pathways three-dimensionally by placing temperature sensors at different depths. By applying AI algorithms, they succeeded in separating and extracting both the depth of blood vessels and the actual blood flow velocity in real time from complex body temperature distributions. Through AI-based analysis, the system can accurately distinguish between vascular depth and actual blood flow velocity within complex temperature patterns of the body.
Experimental results showed that the system successfully measured blood flow velocities in the range of 1–10 mm/s with an error within 0.12 mm/s, and blood vessel depths in the range of 1–2 mm with an error within 0.07 mm. This level of error is smaller than the thickness of a human hair and represents a degree of precision that is difficult to achieve with typical wearable devices.
In particular, when this technology is combined with photoplethysmography (PPG) sensors used in smartwatches, the error in blood pressure measurement can be reduced by up to 72.6%. This indicates that smartwatch-based blood pressure measurements could become much closer to those obtained with hospital equipment. In other words, this achievement can significantly improve the reliability of wearable devices.
This electronic patch can be used in emergency medical settings to detect changes in a patient’s condition in real time. It may also be applied to personalized health management for patients with hypertension or diabetes and to the early detection of acute warning signs such as shock.
Prof. Kyeongha Kwon stated, “This technology provides a fundamental platform for measuring blood flow and blood pressure more accurately, and when combined with smartwatches, it will elevate the level of everyday health monitoring.”
The study was led by Young Min Sim, an integrated M.S.–Ph.D. student, as the first author. The research results were published on February 6 in the world-renowned journal Science Advances.
Paper title: Deep learning–integrated multilayer thermal gradient sensing platform for real-time blood flow monitoring
DOI: 10.1126/sciadv.aea8902
Meanwhile, this research was supported by the Samsung Advanced Institute of Technology (SAIT), the National Research Foundation of Korea (NRF) Outstanding Young Researcher Program (2022R1C1C1010555), the Regional Innovation Leading Research Center (2020R1A5A8018367), the BK21 FOUR Program, and the Artificial Intelligence Semiconductor Graduate School program funded by the Institute of Information & Communications Technology Planning & Evaluation (IITP).
