Publication: Non-Invasive Blood Glucose Estimation Through Vascular Contraction Signal Analysis
2
0
Issued Date
2025-01-01
Resource Type
Scopus ID
2-s2.0-105004560932
Journal Title
10th International Conference on Digital Arts, Media and Technology, DAMT 2025 and 8th ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering, NCON 2025
Start Page
612
End Page
616
Rights Holder(s)
SCOPUS
Bibliographic Citation
10th International Conference on Digital Arts, Media and Technology, DAMT 2025 and 8th ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering, NCON 2025 (2025) , 612-616
Suggested Citation
Pomarin C., Lakhonphon T., Traisaeng S., Jaroensin P., Tantisatirapong S., Chanwimalueang T. Non-Invasive Blood Glucose Estimation Through Vascular Contraction Signal Analysis. 10th International Conference on Digital Arts, Media and Technology, DAMT 2025 and 8th ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering, NCON 2025 (2025) , 612-616. 616. doi:10.1109/ECTIDAMTNCON64748.2025.10962123 Retrieved from: https://hdl.handle.net/20.500.14740/20090
Author's Affiliation
Corresponding Author(s)
Other Contributor(s)
Abstract
Diabetes is a critical health concern globally, as it impacts overall health and serves as a precursor to other conditions. Regular monitoring of blood glucose levels is essential for effective diabetes management. However, current glucose monitoring devices are invasive, leading to tissue damage, pain, and an increased risk of infection. This paper presents a continued study on the development of a non-invasive blood glucose estimation system. Using transmitting light energy in the near-infrared wavelength and near-red wavelength to detect and record PPG signals using an ESP32 microcontroller with a MAX30102 sensor. The MAX30102 sensor can transmit and receive light in red (660 nm) and near-infrared (880 nm) and measure on the fingertip. PPG signals are then analyzed for their prominent features, revealing the relationship between blood glucose levels and the most prominent features, namely root mean square of successive differences (RMSSD) power in the high frequency range (PowHF) and ratio of low frequency range to high frequency range (LF/HF) of the red wavelength. Subsequently, these features were utilized using Multiple regression equations, with one and two variables. Each of 10 Subject participated experimental session lasts approximately 2 hours and 35 minutes. Finally, the best-performing model is the two-variable regression equation model for data from the second segment after signal filtering consisting of Glucose, RMSSD and PowHF of the red wavelength, achieving an accuracy of 89.86%.