Studi Eksperimental Deteksi Kelelahan pada Pengemudi Mobil Berbasis Embedded Fuzzy System

Niate, Oyashi (2025) Studi Eksperimental Deteksi Kelelahan pada Pengemudi Mobil Berbasis Embedded Fuzzy System. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 5023211017-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy

Abstract

Kasus kecelakaan lalu lintas di Indonesia pada 2023 meningkat dengan 94,71% disebabkan kelalaian pengemudi, terutama faktor kelelahan. Penelitian ini mengembangkan sistem deteksi kelelahan pengemudi berbasis Embedded Fuzzy System untuk mengatasi keterbatasan penelitian sebelumnya yang belum optimal dalam instrumentasi ECG, tidak compact, dan sistem fuzzy konvensional yang kurang efisien. Sistem mengintegrasikan tiga variabel, yaitu variabel subjek (heart rate dari ECG dengan Conductive Plate Electrode dan data postural dari sensor IMU), variabel kendaraan (throttle position melalui modul OBD-II Freematics), dan variabel lingkungan (kecepatan dan durasi berkendara dari GPS yang diolah menggunakan Haversine Formula). Instrumentasi ECG menggunakan filter Electromagnetic Interference (EMI) berupa pelindung aluminium untuk mengurangi noise dari mesin kendaraan. Sistem menggunakan arsitektur master-slave dengan STM32F411CEU6 untuk variabel subjek dan Arduino Uno untuk variabel kendaraan dan lingkungan sebagai slave, serta Raspberry Pi 4b sebagai modul master. Implementasi Hierarchical Fuzzy System dua level meningkatkan efisiensi komputasi dengan lower level fuzzy pada masing-masing slave mengklasifikasikan tingkat kelelahan dan keamanan berkendara, kemudian upper level fuzzy pada master mengintegrasikan hasil untuk klasifikasi final keselamatan berkendara (Safe, Warning1, Warning2, Danger). Hasil pengujian menunjukkan Conductive Plate Electrode mencapai SNR 12,71 dB, algoritma Pan-Tompkins efektif mendeteksi QRS complex meski terdapat motion artifact, dan GPS dengan Kalman Filter memberikan koordinat meskipun beberapa kali mengalami signal loss di area tertutup. Evaluasi sistem fuzzy pada simulator (4 subjek, 48 sampel) mencapai accuracy 70,8%, sementara pengujian mobil (3 subjek, 36 sampel) menghasilkan accuracy 83,3%. Dari pengujian juga diperoleh adanya bias pada sistem deteksi untuk kondisi transisi, yaitu Drowsy1 dan Drowsy1. Terdapat masalah kompatibilitas pada modul OBD-II, dimana hanya 1 dari 4 mobil yang diuji dapat terhubung dan diakuisisi datanya, akibat evolusi protokol ECU pada mobil keluaran terbaru. Sebagai solusi adaptif, dilakukan modifikasi sistem menggunakan GPS untuk analisis data perilaku berkendara.
==================================================================================================================================
Traffic accidents in Indonesia increased in 2023 with 94.71% caused by driver negligence, primarily due to fatigue. This research develops an Embedded Fuzzy System-based driver fatigue detection system to address previous study limitations including ECG instrumentation, non-compact design, and inefficient conventional fuzzy systems. The system integrates three variables, consists of subject variables (heart rate from ECG using Conductive Plate Electrodes and postural data from IMU sensors), vehicle variables (throttle position via Freematics OBD-II module), and environmental variables (driving speed and duration from GPS processed using Haversine Formula). ECG instrumentation uses Electromagnetic Interference (EMI) filter with aluminum shielding to reduce noise from vehicle engines. The system employs master-slave architecture with STM32F411CEU6 for subject variables and Arduino Uno for vehicle-environmental variables as slaves, and Raspberry Pi 4b as master controller. Two-Level Hierarchical Fuzzy System implementation improves computational efficiency with lower level fuzzy on each slave classifying fatigue levels and driving safety, then upper level fuzzy on master integrates results for final driving safety classification (Safe, Warning1, Warning2, Danger). Test results show Conductive Plate Electrodes achieve 12.71 dB SNR, Pan-Tompkins algorithm effectively detects QRS complex despite motion artifacts, and GPS with Kalman Filter provides accurate coordinates despite signal loss in covered areas. Fuzzy system evaluation on simulator (4 subjects, 48 samples) achieved 70.8% accuracy, while vehicle testing (3 subjects, 36 samples) achieved 83.3% accuracy. From the testing, it was also found that there is bias in the detection system for transitional conditions, namely Drowsy1 and Drowsy2. There are compatibility issues with the OBD-II module, where only 1 out of 4 tested vehicles could be connected and have their data acquired, due to the evolution of ECU protocols in newer vehicle models. As an adaptive solution, system modification was implemented using GPS for driving behavior data analysis.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Deteksi kelelahan pengemudi, sistem tertanam, ECG, pengambilan keputusan fuzzy hierarkis, keselamatan berkendara, perilaku mengemudi, Driver fatigue detection, embedded system, hierarchical fuzzy decision support system, safety driving
Subjects:	H Social Sciences > HE Transportation and Communications > HE5614.2 Traffic safety Q Science > QA Mathematics > QA9.64 Fuzzy logic T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL798.N3 Global Positioning System.
Divisions:	Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Biomedical Engineering > 11410-(S1) Undergraduate Thesis
Depositing User:	Oyashi Bunge Niate
Date Deposited:	04 Aug 2025 03:25
Last Modified:	04 Aug 2025 03:25
URI:	http://repository.its.ac.id/id/eprint/126385

Actions (login required)

View Item