Challenges in detecting drowsiness based on driver's behavior

Drowsiness while driving has been a critical issue within the context of transportation safety. A number of approaches have been developed to reduce the risks of drowsy drivers. The mechanisms in detecting fatigue and sleepiness while driving has been categorized into three broad approaches, including vehicle-based, physiological-based, and behavior-based approaches. This paper will discuss recent studies in recognizing drowsy drivers based on their behaviors, particularly changes in eyes and facial characteristics. This paper will also address challenges in capturing aspects of natural expressions, driver responses, behavior, and task environment associated with sleepiness. Additionally, a number of technical aspects should be seriously considered, including correctly capturing face and eye characteristics from unwanted movements, unsuitable task environments, technological limitations, and individual differences. (paper)