Continuous non-invasive blood glucose level measurement using near-infrared LEDs

• Main Authors: Yaacob, Arif Amiruddin, Abdul Malik, Noreha, Alam, A. H. M. Zahirul, Salim, Siti Nur Atikah
• Format: Conference or Workshop Item
• Language: English; English; English
• Published: IEEE 2021
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://irep.iium.edu.my/91406/6/91406_Continuous%20Non-Invasive%20Blood%20Glucose%20_abstract%20book.pdf; http://irep.iium.edu.my/91406/7/91406_Continuous%20Non-Invasive%20Blood%20Glucose%20_comf%20article.pdf; http://irep.iium.edu.my/91406/18/91406_Continuous%20Non-Invasive%20Blood%20Glucose%20Level%20Measurement%20using%20Near-Infrared%20LEDs_Scopus.pdf; http://irep.iium.edu.my/91406/; https://ieeexplore-ieee-org.ezlib.iium.edu.my/document/9467165

Constant blood glucose level monitoring is important especially to diabetic patients. However, the current methods available are invasive in nature and cause pain while also being costly. In addition, there is also the risk of contracting infections with improper hygienic practice. Therefore, in this project, a near-infrared (NIR) based LED sensor circuit has been developed to achieve a non-invasive method as an alternative for measuring blood glucose level. The system used 940 nm, 1450 nm, and 1550 nm wavelength NIR LEDs. The emitters and receiver were placed in a black box to reduce noise interference. The analog signal was transmitted through a filter and an amplifier circuit. Then, it was converted to a digital signal where the output voltage was measured, and the glucose level was calculated using an Arduino UNO. In the in vitro test, it was observed that the output voltage changes with changing concentrations of glucose solution, and there was a good correlation between the two variables, and this motivated the in vivo test. Measurement in transmittance yielded a 0.74 for coefficient of determination and 0.72 for reflectance. It was found that the combination of two spectroscopic measurement techniques improved the root mean square error to 16.02 mg/dL and reduced the percentage error to 9.56% compared to individual results. Thus, this non-invasive method can be an alternative for current invasive glucose measuring devices.