Identification of Blood Sugar Based on Non-invasive Measurements UsingPhotoplethysmography Method Signal Decoding in Diabetics in Tasikmalaya

Abstract

Non-invasive blood sugar testing tools are now in great demand among the public. This research aims to develop a non-invasive blood sugar examination tool with a Photopletysmograph signal decoding system based on data acquisition from a wearable sensor device on the wrist. The tool used for data acquisition uses Near Infra Red Light-emitting Diode NIR LED 880 nm, 660 nm and 537 nm. This acquisition transfers signals from the body via a photodiode to a PC Personal Computer for processing using Matlab. The method used in this research is to look for the height of the systolic and diastolic amplitude of the Photopletysmograph PPG signal and use the peak-to-peak voltage Vpp to implement Beer Lambert's law. This method was tested on 54 people, randomly aged 26-89 years, with normal blood sugar to high blood sugar. The results show an error of 18.37% from the gold standard. In conclusion, the use of the PPG signal decoding method has proven to be significant in producing blood sugar values. Furthermore, this method was developed for real-time remote measurements via wearable sensor devices.