Design and Development of a Non-Invasive Blood Glucose Level Measurement Device Based on Arduino Uno and Near-Infrared (NIR) Sensor

Diabetes Mellitus is one of the leading causes of death in the world. Therefore, regular monitoring of blood sugar levels is very important to prevent complications. However, monitoring blood sugar using conventional-invasive methods can cause discomfort and risk causing infection. This study aims to design a non-invasive blood sugar level measuring device based on Arduino Uno R3 using a BPW34 Near-Infrared (NIR) sensor reinforced with an LM358 operational amplifier. The measurement results are displayed on a 16x2 LCD and sent via Telegram Bot in real-time. This study focused on Type 2 Diabetes Mellitus and was tested on 10 participants aged 18-45 years with some patients in fasting and some not fasting conditions. The test results showed good accuracy with a Mean Absolute Error of 2.48%, minimal systematic bias (-0.08%), and a very strong correlation coefficient (r = 0.996436) against invasive methods. All measurement data meets the tolerance criteria of ISO 15197:2013 with an error range of -4.55% to +3.16%, indicating that this tool has the potential to be a more convenient and safer alternative for initial blood sugar level screening