Pengembangan Ambulatory Cardiac Monitor Untuk Deteksi Aritmia Secara Kontinu Menggunakan Artificial Neural Network

Harza, Najla Rasikha Putri (2023) Pengembangan Ambulatory Cardiac Monitor Untuk Deteksi Aritmia Secara Kontinu Menggunakan Artificial Neural Network. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 07311940000044-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 October 2025. Download (7MB) | Request a copy

Abstract

Aritmia didefinisikan sebagai detak jantung yang tidak teratur. Pada praktik medis, deteksi aritmia bergantung pada hasil wawancara pasien dan analisis secara visual terhadap sinyal ECG yang dilakukan oleh tenaga ahli medis. Metode alternatif lain adalah penggunaan monitor ECG ambulatory yang dapat merekam sinyal jantung dalam waktu yang lama untuk mendeteksi aritmia, yaitu monitor Holter. Namun, dalam penggunaannya, Holter hanya dapat merekam sinyal jantung yang kemudian dianalisis manual oleh dokter. Holter tidak dapat mengekstraksi fitur sinyal jantung maupun mendeteksi aritmia secara otomatis. Pengembangan sistem deteksi aritmia yang telah ada menggunakan algoritma yang cukup kompleks, sehingga daya komputasi dan memori yang dibutuhkan besar, dan tidak sesuai diterapkan pada alat yang bersifat wearable karena memiliki kemampuan komputasi dan memori yang terbatas. Maka dari itu pada penelitian ini dikembangkan alat monitor sinyal jantung ambulatory yang dapat mendeteksi aritmia jantung secara otomatis serta bersifat wearable dengan beban komputasi yang rendah. Pada penelitian ini digunakan metode feature extraction aritmia PVC dengan beban komputasi rendah agar dapat berjalan di mikrokontroler STM32. Mikromedia 5 for STM32F4 digunakan sebagai development board yang diberikan power supply eksternal agar dapat digunakan dalam waktu yang lama. Jenis aritmia yang menjadi objek penelitian adalah Premature Ventricular Contraction (PVC). PVC disebabkan oleh depolarisasi miokardium prematur di daerah ventrikel yang menyebabkan disrupsi pada aktivitas elektrik dan mekanik jantung. Artificial neural network jenis multilayer perceptron (MLP) digunakan sebagai metode klasifikasi untuk mendeteksi detak jantung PVC dari detak jantung normal. Diperoleh hasil akurasi 97,12%, presisi 98,07%, sensitivitas 96,13%, dan spesifisitas 98,1%. Algoritma deteksi PVC berhasil ditanamkan pada mikrokontroler dengan menggunakan 25% dari keseluruhan flash memory. Diuji coba penggunaan powerbank sebagai power supply dengan durasi ketahanan sekitar 25 jam. Hasil beat PVC yang terdeteksi dan informasi waktu terjadinya dapat disimpan pada SD card dengan memanfaatkan penerapan RTC.

===================================================================================================================================

Arrhythmia is defined as irregular heartbeats. In medical practice, the detection of arrhythmias relies on the results of patient interviews and visual analysis of ECG signals performed by medical professionals. Another alternative method is the use of an ambulatory ECG monitor that can record heart signals for a long time to detect arrhythmias, namely the Holter monitor. However, in use, Holter can only record heart signals which are then manually analyzed by doctors. Holter can neither extract cardiac signal features nor detect arrhythmias automatically. The development of an existing arrhythmia detection system uses a fairly complex algorithm, so that the computational power and memory required are large, and it is not suitable to be applied to wearable devices because they have limited computing and memory capabilities. Therefore, in this study, an ambulatory heart signal monitoring device was developed which can detect cardiac arrhythmias automatically and is wearable with a low computational load. In this research, the PVC arrhythmia feature extraction method is used with a low computational load so that it can run on the STM32 microcontroller. Micromedia 5 for STM32F4 is used as a development board which is provided with an external power supply so that it can be used for a long time. The type of arrhythmia that is the object of research is Premature Ventricular Contraction (PVC). PVCs are caused by premature myocardial depolarization in the ventricular region causing a disruption in the electrical and mechanical activity of the heart. Multilayer perceptron (MLP) artificial neural network is used as a classification method to detect PVC heartbeats from normal heartbeats. The results obtained were 97.12% accuracy, 98.07% precision, 96.13% sensitivity, and 98.1% specificity. The PVC detection algorithm was successfully embedded in the microcontroller using 25% of the total flash memory. Tested using a power bank as a power supply with a duration of around 25 hours. The results of detected PVC beats and information on the time of occurrence can be stored on an SD card by implementing RTC.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Aritmia, PVC, ECG, ambulatory, wearable, STM32, Mikromedia 5, feature extraction, ANN ============================================================ Arrhythmia, PVC, ECG, ambulatory, wearable STM32, Mikromedia 5, feature extraction, ANN
Subjects:	T Technology > T Technology (General) T Technology > TK Electrical engineering. Electronics Nuclear engineering T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7871.674 Detectors. Sensors T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7882.B56 Biometric identification
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Najla Rasikha Putri Harza
Date Deposited:	27 Jul 2023 13:38
Last Modified:	27 Jul 2023 13:38
URI:	http://repository.its.ac.id/id/eprint/100033

Actions (login required)

View Item