Design and Implementation of a Video Oculography-based Hybrid Driver Drowsiness and Over speeding Detection System

Abstract

Driver drowsiness and over speeding are major contributors to road accidents worldwide, leading to death and serious injuries. This paper presents a real-time drowsiness and over speedingdetection. The drowsiness detection method uses a Python-based algorithm to calculate the Eye Aspect Ratio (EAR),Mouth Aspect Ratio (MAR) and measure head tilt from video frames, providing nonintrusive driver monitoring system. The vehicle speed is acquired through an onboard sensor integrated within the vehicle, and sensor emulation is employed for data acquisition, analysis, and validation of the proposed system.. The system integrates audio-visual warnings and sends alerts to a mobile application to inform the driver. Experimental results demonstrate an accuracy of approximately 92% in drowsiness detection and reliable over speedingalerting, validating the system’s effectiveness in enhancing vehicular safety. Challenges in lighting and occlusions alongside future directions including multimodal sensing and IoT integration are discussed.