Efficient driver drowsiness detection at moderate levels of drowsiness.

Area of Science:

• Human Factors Engineering
• Automotive Safety
• Cognitive Psychology

Background:

• Existing driver drowsiness detection methods often rely on lane deviation and high fatigue levels.
• There is a need for reliable detection of drowsiness at moderate fatigue levels, before accident risk increases.

Purpose of the Study:

• To develop and evaluate a novel method for detecting driver drowsiness at moderate fatigue levels.
• To identify reliable driving metrics for early-stage drowsiness detection.

Main Methods:

• Evaluated 87 drowsiness detection metrics in simulated shift work studies with high-fidelity driving simulators.
• Utilized principal component analysis to reduce data dimensionality and identify dominant driving metrics.
• Correlated driving metrics with an independent measure of fatigue (psychomotor vigilance test).

Main Results:

• Steering wheel variability and lateral lane position variability emerged as two dominant dimensions.
• Lateral lane position variability showed the strongest correlation with fatigue levels.
• Lateral lane position variability can be derived from steering wheel angle changes, offering an alternative to video-based systems.

Conclusions:

• Steering wheel variability is a promising indicator for detecting driver drowsiness at moderate fatigue levels.
• This method provides a cost-effective and easily installable alternative to traditional video-based drowsiness detection systems.
• The findings contribute to developing advanced driver assistance systems for improved road safety.