Sistem Prediksi Perawatan Kendaraan Berbasis On Board Diagnostics Dan Neural Network

Vauji, Bara Satria Sakti Raharja (2026) Sistem Prediksi Perawatan Kendaraan Berbasis On Board Diagnostics Dan Neural Network. Masters thesis, Institut Teknologi Sepuluh Nopember.

Text 6022242012_Master_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Perkembangan teknologi otomotif menuntut sistem pemantauan kondisi kendaraan yang tidak hanya mampu membaca data operasional secara real-time, tetapi juga mendukung prediksi kebutuhan perawatan berdasarkan kondisi aktual kendaraan. Penelitian ini mengembangkan sistem prediksi perawatan kendaraan berbasis On-Board Diagnostics (OBD-II) dan Neural Network dengan memanfaatkan Raspberry Pi sebagai perangkat edge computing. Data kendaraan diperoleh melalui modul OBD-II ELM327 yang terhubung ke Electronic Control Unit (ECU), kemudian digunakan untuk mengakuisisi parameter Engine RPM, Throttle Position Sensor (TPS), dan Engine Coolant Temperature (ECT) secara real-time. Dataset yang diperoleh diproses melalui tahap preprocessing dan digunakan untuk melatih tiga model deep learning, yaitu Artificial Neural Network (ANN), Convolutional Neural Network (CNN), dan Long Short-Term Memory (LSTM), untuk mengklasifikasikan kondisi kendaraan ke dalam kategori Normal, Warning, dan Urgent. Evaluasi model dilakukan menggunakan metrik accuracy, precision, recall, F1-score, kurva pelatihan, dan confusion matrix. Hasil penelitian menunjukkan bahwa sistem monitoring berbasis OBD-II dan edge computing berhasil diimplementasikan untuk memantau kondisi kendaraan secara real-time. Ketiga model mampu melakukan klasifikasi kondisi kendaraan dengan tingkat akurasi di atas 83%, dengan CNN memberikan performa terbaik sebesar 83,89%, diikuti oleh LSTM sebesar 83,56% dan ANN sekitar 83%. Hasil tersebut menunjukkan bahwa kombinasi teknologi OBD-II dan metode deep learning mampu mendukung implementasi predictive maintenance sehingga dapat membantu mendeteksi potensi gangguan kendaraan secara lebih dini dan meningkatkan efektivitas proses perawatan kendaraan.
======================================================================================================================================
The rapid advancement of automotive technology has increased the demand for vehicle condition monitoring systems that are not only capable of acquiring real time operational data but also able to predict maintenance requirements based on the actual condition of the vehicle. This study develops a Vehicle Maintenance Prediction System Based on On-Board Diagnostics (OBD-II) and Neural Networks by utilizing a Raspberry Pi as an edge computing platform. Vehicle data are acquired through an OBD-II ELM327 module connected to the Electronic Control Unit (ECU) to collect real-time engine parameters, including Engine Revolutions Per Minute (RPM), Throttle Position Sensor (TPS), and Engine Coolant Temperature (ECT). The acquired dataset is preprocessed and used to train three deep learning models, namely Artificial Neural Network (ANN), Convolutional Neural Network (CNN), and Long Short-Term Memory (LSTM), to classify vehicle conditions into three categories: Normal, Warning, and Urgent. Model performance is evaluated using accuracy, precision, recall, F1-score, training curves, and confusion matrix analysis. The results demonstrate that the proposed OBD-II and edge computing-based monitoring system successfully performs real time vehicle condition monitoring. All three deep learning models achieve classification accuracies above 83%, with the CNN model providing the best performance, achieving an accuracy of 83.89%, followed by LSTM with 83.56% and ANN with approximately 83%. These findings indicate that the integration of OBD-II technology and deep learning methods can effectively support predictive maintenance, enabling early detection of potential vehicle faults and improving the effectiveness and reliability of vehicle maintenance.

Item Type:	Thesis (Masters)
Uncontrolled Keywords:	Electronic Control Unit, On-Board Diagnostics, dan Neural Network.Electronic Control Unit, On-Board Diagnostics, and Neural Network
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7870.23 Reliability. Failures T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7872 Electromagnetic Devices T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7878 Electronic instruments
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20101-(S2) Master Thesis
Depositing User:	Bara Satria Sakti Raharja Vauji
Date Deposited:	28 Jul 2026 02:43
Last Modified:	28 Jul 2026 02:43
URI:	http://repository.its.ac.id/id/eprint/137692

Actions (login required)

View Item