Real-time driver drowsiness and distraction detection using convolutional neural network with multiple behavioral features

ABSTRACT: Road accidents caused by driver drowsiness and distraction represent significant threats to worldwide road safety, with facilities and injuries at alarming rates in the Philippines. With asignificant amount of casualties, the need for proactive measures is urgent. Recognizing the human factor as the primary cause of accidents, this study aimed to develop a real-time driver drowsiness and distraction detection system to mitigate risks. Using non-intrusive camera sensors and convolutional neural networks (CNN), the system monitors the driver behaviour, including facial expressions, eye movements, and lane position, to detect signs of drowsiness and distraction. This study meticulously outlines the systematic procedures, employing a quantitative developmental research approach to design and assess the effectiveness of the system. Real-world on-road testing with participants engaged in long-duration driving ensures the authenticity of data collection. The findings highlight the sytem’s prmising performance in drowsiness and distraction detection, with high accuracy rates and effective alert system triggered upon detection of potential risks. The integration of CNN technology underscores the system’s potential to significantly enhance road safety authorities. This research sets a foundation for future advancements in proactive driver safety technologies, emphasizing the critical importance of addressing driver drowsiness and distraction on the roads.

ABSTRACT:

Road accidents caused by driver drowsiness and distraction represent significant threats to worldwide road safety, with facilities and injuries at alarming rates in the Philippines. With asignificant amount of casualties, the need for proactive measures is urgent. Recognizing the human factor as the primary cause of accidents, this study aimed to develop a real-time driver drowsiness and distraction detection system to mitigate risks. Using non-intrusive camera sensors and convolutional neural networks (CNN), the system monitors the driver behaviour, including facial expressions, eye movements, and lane position, to detect signs of drowsiness and distraction. This study meticulously outlines the systematic procedures, employing a quantitative developmental research approach to design and assess the effectiveness of the system. Real-world on-road testing with participants engaged in long-duration driving ensures the authenticity of data collection. The findings highlight the sytem's prmising performance in drowsiness and distraction detection, with high accuracy rates and effective alert system triggered upon detection of potential risks. The integration of CNN technology underscores the system's potential to significantly enhance road safety authorities. This research sets a foundation for future advancements in proactive driver safety technologies, emphasizing the critical importance of addressing driver drowsiness and distraction on the roads.

Filipiniana