A feasability study of drowsiness detection using driving behaviour parameters

One of the main causes of car accidents is drowsiness. There have been many studies regarding driving monitoring systems in the past few years, although most of them are focused in simulator environments. This paper presents a system to detect drowsiness patterns in real driving environments, where many external conditions need to be taken into account. Initial tests were done in simulator, followed by tests in real vehicles. Although two different approaches have been developed, this paper is focused in the inadequate driving identification based on the steering movements. Its sub-modules are also presented, with a special focus on the active driving detector.