Ontology-Based Approach and Implementation of ADAS System for Mobile Device Use While Driving

Abstract

Many people drive while being tired or drowsy and according to experts, many drivers fail to recognize they are in a fatigued state. The paper describes an ontology-based approach and a real-time prototype of a computer vision system for monitoring driver’s dangerous behaviour patterns such as distraction, drowsiness and reduced vigilance. Our approach focuses on widely used mobile devices, such as a smartphone or a tablet, which have become an integral part of our daily life. The main components of the system consist of a front-facing camera, gyroscope sensor installed on the smartphone and various computer vision algorithms for simultaneous, real-time and non-intrusive monitoring of various visual bio-behaviours that typically characterize a driver’s level of vigilance. The visual behaviours include eyelid movement (PERCLOS percentage of eye closure, eye blink rate), face orientation (face pose), and gaze movement (pupil movement). The two ontological models presented in the paper are a driver model and a vehicle model. Based on these models and the information available from cameras and sensors, the context is created that allows the system to determine dangerous situations using the driver mobile device mounted on the windshield of the vehicle. The system was tested in a simulated environment with subjects of different ethnic backgrounds, genders, ages, with/without glasses, and under different illumination conditions, and it was found very robust, reliable and accurate.