ABSTRACT

This study aimed to develop a fatigue driving monitoring system based on data mining and visualization technology to improve road safety. The system collects drivers’ facial feature data, such as blink rate, yawning, and head posture, and utilizes machine learning algorithms to detect and warn about fatigue in real time. First, it cleans and extracts a substantial amount of driving data, then trains a fatigue detection model that accurately identifies the driver’s level of fatigue. Additionally, the system features a visualization function that displays the driver’s physiological state and driving behavior, helping the driver to understand their fatigue status promptly. Moreover, based on the identified fatigue level, the system provides tailored tips and suggestions, such as recommending appropriate rest or adjusting the driving rhythm, to mitigate the risk of fatigued driving. Based on the evaluation of the IT experts, the developed system complied with ISO 25010 Software Quality Standards to a very great extent. Furthermore, the system has demonstrated significant effectiveness in improving driving safety and reducing traffic accidents. This study offers valuable practical insights into the development of intelligent transportation systems.

Keywords: Convolutional neural networks, fatigue driving, fatigue driving detection, SSD
https://doi.org/10.57180/hplx8173