Modeling drivers' reaction when being tailgated: A Random Forests Method

Area of Science:

• Traffic Safety
• Human Factors in Driving
• Aggressive Driving Behavior

Background:

• Tailgating is a major cause of rear-end collisions.
• Research has explored tailgating drivers' behavior and solutions.
• This study focuses on highway tailgating scenarios and leading vehicle reactions.

Purpose of the Study:

• To understand highway tailgating scenarios.
• To examine leading vehicles' reactions to tailgating.
• To identify factors influencing these reactions.

Main Methods:

• Identified 1,255 tailgating events using a one-second time headway threshold.
• Classified leading vehicle reactions into four types: lane change, slowing, speeding up, or no response.
• Employed a Random Forests algorithm to predict reactions based on driver, vehicle, and environmental variables.

Main Results:

• Tailgating is more prevalent among male drivers, in sunny weather, and during daytime.
• Lane changing was the most common leading vehicle reaction (over 50%).
• Mean time headway, duration, and minimum time headway significantly impacted leading vehicle reactions.

Conclusions:

• Leading vehicles change lanes in 95% of events when tailgated for two minutes or longer.
• Findings enhance understanding of driver behavior and decision-making.
• Results can inform the design of countermeasures and assistance systems to mitigate tailgating and its safety consequences.