Assessing driver behavior in work zones: A discretized duration approach to predict speeding

Area of Science:

• Traffic Safety Engineering
• Transportation Science
• Predictive Analytics

Background:

• High speeds in work zones increase crash likelihood and severity.
• Reduced speed limits aim to improve traffic flow but can cause abrupt speed changes, increasing rear-end crash risks.
• Driver non-compliance with work zone speed limits is a persistent safety challenge.

Purpose of the Study:

• To develop and validate a duration-based prediction framework for forecasting speeding occurrences in highway work zones.
• To identify key predictors influencing speeding behavior within work zones.
• To enhance proactive safety measures and enforcement strategies in work zones.

Main Methods:

• Utilized a duration-based prediction framework to forecast speeding events.
• Identified significant predictors including visibility, number of lanes, posted speed limit, segment length, speed variation, and travel time index.
• Validated model accuracy based on the duration between consecutive speeding events.

Main Results:

• Number of lanes, posted speed limit, and speed variation positively correlate with speeding.
• Visibility, segment length, and travel time index negatively correlate with speeding.
• Model accuracy is higher for more frequent speeding events, predicting within 61% of actual occurrences for events within a 5-hour interval.

Conclusions:

• The duration-based prediction framework effectively forecasts speeding in work zones, particularly where speeding is frequent.
• Proactive prediction enables targeted speed enforcement, enhancing work zone safety for road users and personnel.
• This predictive model serves as a valuable tool for agencies to manage and mitigate speeding-related hazards in real-time.