Sistem Pemantauan Segmen ST Dengan Elektroda Non-Kontak Elektrokardiogram Untuk Deteksi ST-Elevation Myocardial Infarction Bagi Pengendara Mobil

Fleming, Alexander (2025) Sistem Pemantauan Segmen ST Dengan Elektroda Non-Kontak Elektrokardiogram Untuk Deteksi ST-Elevation Myocardial Infarction Bagi Pengendara Mobil. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5023211054-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (11MB) | Request a copy

Abstract

Penelitian ini bertujuan untuk mengembangkan sistem deteksi dini ST-Elevation Myocardial Infarction (STEMI) pada pengendara mobil menggunakan sensor capacitive ECG yang terintegrasi pada kursi pengemudi. Sistem prototipe fungsional ini menerapkan arsitektur Internet of Things (IoT) untuk transmisi data nirkabel melalui protokol MQTT, memungkinkan analisis sinyal secara real-time dan pemberian peringatan dini. Keunggulan sensor kapasitif adalah kenyamanan dan kemampuan akuisisi non-kontak tanpa membutuhkan gel konduktif. Sinyal EKG yang didapatkan diproses melalui modul front-end EKG (MikroElektronika ECG Click) yang sudah terintegrasi untuk penguatan sinyal dan filterisasi, serta filter digital (bandpass Butterworth orde 4 dengan frekuensi cutoff 0.5Hz dan 40Hz) untuk mengatasi noise dan artefak. Deteksi fitur fisiologis EKG, seperti puncak R, titik isoelektrik, dan J-Point, dilakukan dengan akurasi tinggi (Puncak R: 96.68%, P-Wave: 95.83%, T-Wave: 96.17%). Dua metode klasifikasi diimplementasikan dan dievaluasi: metode berbasis aturan fisiologis (rule-based thresholding) dan model machine learning berbasis GRU. Hasil komparatif menunjukkan bahwa metode berbasis aturan fisiologis lebih unggul dalam ketahanan terhadap noise dunia nyata, efisiensi komputasi, dan transparansi logika, menjadikannya ideal untuk implementasi pada perangkat embedded. Validasi sistem terhadap sinyal EKG standar medis menunjukkan kemiripan morfologis dan korelasi linear yang tinggi (70%-90%), membuktikan kelayakan pendekatan akuisisi kapasitif untuk aplikasi ini meskipun dihadapkan pada tingkat derau yang lebih tinggi. Pengujian spesifisitas pada sinyal patologis non-STEMI menunjukkan hasil garis ST-Elevation sebesar ±78,70mm, menegaskan keandalan dan spesifisitas klinisnya sebagai alat peringatan dini tanpa memicu false alarm.
==============================================================
This research aims to develop an early detection system for ST-Elevation Myocardial Infarction (STEMI) in car drivers using a capacitive ECG sensor integrated into the driver's seat. This functional prototype system implements an Internet of Things (IoT) architecture for wireless data transmission via the MQTT protocol, enabling real-time signal analysis and the delivery of early warnings. The advantages of the capacitive sensor are its comfort and non-contact acquisition capability, without requiring conductive gel. The acquired ECG signal is processed through an integrated ECG front-end module (MikroElektronika ECG Click) for signal amplification and filtering, as well as a digital filter (a 4th-order Butterworth bandpass filter with cutoff frequencies of 0.5Hz and 40Hz) to mitigate noise and artifacts. The detection of physiological ECG features, such as the R-peak, isoelectric point, and J-Point, was performed with high accuracy (R-Peak: 96.68%, P-Wave: 95.83%, T-Wave: 96.17%). Two classification methods were implemented and evaluated: a physiology rule-based method (rule-based thresholding) and a GRU-based machine learning model. Comparative results show that the rule-based method is superior in terms of robustness against real-world noise, computational efficiency, and logical transparency, making it ideal for implementation on embedded devices. System validation against standard medical ECG signals showed morphological similarity and a high linear correlation (70%-90%), proving the feasibility of the capacitive acquisition approach for this application, despite facing higher noise levels. Specificity testing on non-STEMI pathological signals showed an ST-Elevation result of ±78.70mm, confirming its reliability and clinical specificity as an early warning tool without triggering false alarms.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Elektroda Non-Kontak, EKG, Machine Learning, Pemantauan Jantung, STEMI, Non-Contact Electrodes, ECG, Heart Monitoring.
Subjects:	H Social Sciences > HE Transportation and Communications > HE5614.2 Traffic safety Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. R Medicine > RC Internal medicine > RC683.5.E5 Electrocardiography T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5102.9 Signal processing.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Alexander Fleming
Date Deposited:	04 Aug 2025 06:21
Last Modified:	04 Aug 2025 06:21
URI:	http://repository.its.ac.id/id/eprint/126517

Actions (login required)

View Item