Sistem Estimasi Kadar Glukosa Darah Non-Invasif Berbasis Multi-Channel Photoplestysmography (PPG)

Manik, Ni Putu Wira Sekar (2025) Sistem Estimasi Kadar Glukosa Darah Non-Invasif Berbasis Multi-Channel Photoplestysmography (PPG). Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5023211041-Undergraduated_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (5MB) | Request a copy

Abstract

Penelitian akan menjelaskan pengembangan sistem berbasis multiple-channel Photoplethysmography (PPG) untuk mendeteksi kadar glukosa darah secara non-invasif. Kebutuhan sistem ini muncul akibat meningkatnya kasus diabetes melitus yang dipicu oleh gaya hidup modern yang tidak sehat dan rendahnya kesadaran masyarakat melakukan langkah preventif dan pemantauan rutin untuk menjaga kadar gula darah. Metode tradisional invasif yang tidak nyaman mendorong pengembangan alat ukur alternatif non-invasif. Rancangan alat ini menggunakan teknologi Photoplethysmography (PPG) dengan 2 kanal sensor cahaya yang dikumpulkan dari 2 jenis LED yaitu merah dan infrared. Fitur sinyal PPG yang diekstrak dikorelasikan dengan kadar glukosa subjek diolah dengan machine learning hingga menghasilkan estimasi kadar glukosa darah. Algoritma machine learning menganalisis data lebih lanjut, menunjukkan korelasi antara konsentrasi glukosa dalam darah dan amplitudo sinyal. Model regresi berbasis decision tree dikonstruksi, dan digunakan algoritma XGBoost (XGB) untuk mencapai nilai estimasi prediksi kadar glukosa berdasarkan real-time samples dengan perbandingan menggunakan glukometer invasif sampel darah. Untuk memastikan akurasi tinggi, sistem ini mengevaluasi metrik seperti Mean Absolute Error (MAE), Root-Mean-Squared Error (RMSE), Pearsons r dan melakukan Clarke Error Grid Analysis (CEGA). Penelitian diharapkan menghasilkan alat portable berukuran kecil.
===================================================================================================================================
The project will describe the development of a multiple-channel photoplethysmography (PPG)-based system for non-invasive blood glucose detection. This system arises from the increasing cases of diabetes mellitus, driven by unhealthy modern lifestyles and the lack of public awareness in regularly monitoring blood glucose levels. The discomfort of traditional invasive methods encourages the development of alternative non-invasive measurement tools. This device design uses photoplethysmography (PPG) technology with two light sensor channels collected from two types of LEDs. Extracted PPG signal features are correlated with the subjects glucose levels, processed through machine learning to estimate blood glucose levels. Machine learning algorithms further analyze the data, showing a correlation between blood glucose concentration and signal amplitude. A decision tree-based regression model is constructed, using XGBoost (XGB) algorithm to achieve predicted glucose levels based on real-time samples, compared to invasive blood glucose meter samples. To ensure high accuracy, the system evaluates glucose metrics mean Absolute Error (MAE), Root-Mean-Squared Error (RMSE), Pearsons r, and Clarke Error Grid Analysis (CEGA). The research results in a small portable device.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Non-invasif, Photoplethysmography (PPG), Estimasi glukosa, Machine learning, Non-invasive, Photoplethysmography (PPG), Glucose Estimation, Machine Learning
Subjects:	T Technology > T Technology (General) > T57.5 Data Processing T Technology > T Technology (General) > T58.62 Decision support systems T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing. T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7871.674 Detectors. Sensors T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7878 Electronic instruments
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Ni Putu Wira Sekar Manik
Date Deposited:	04 Aug 2025 01:52
Last Modified:	04 Aug 2025 01:52
URI:	http://repository.its.ac.id/id/eprint/126106

Actions (login required)

View Item