A Novel Real-Time Redundant System For Aiding Drivers To Increase Their Attention Level

Road accidents are a major concern worldwide. Research of technological solutions to improve road safety is an open topic, considering this issue from various perspectives. As widely reported in the literature, driver distraction is one of the most common causes of collisions. Even if vehicles are becoming safer and safer over the years, pedestrians and cyclists are still exposed to severe accidents. Major concerns about driver distraction is mobile phone use and drowsy driving. This paper proposes a virtual buddy designed to help drivers, thanks to warning sound messages, improve their attention level. The main idea is to recognize, thanks to different sources (video camera, physiological data, and interior environmental conditions) interpreted by a data-fusion algorithm, whenever there is a distracted or drowsy behavior, recalling the driver to the road. The effectiveness of the data-fusion algorithm in detecting dangerous conditions has been verified thanks to a driving simulator experimentally, obtaining no false negatives from distraction from drowsiness, a sensitivity of about 85% for the distraction caused by mobile phone usage or other activities different from focusing on driving.