Predicting crash-relevant violations at stop sign-controlled intersections for the development of an intersection...

Area of Science:

• Traffic Safety Engineering
• Automotive Systems
• Predictive Modeling

Background:

• Intersection crashes cause thousands of fatalities annually in the US.
• Intersection Advanced Driver Assistance Systems (I-ADAS) aim to enhance safety by detecting potential hazards and assisting drivers.
• Developing predictive models for stop sign violations is crucial for I-ADAS functionality.

Purpose of the Study:

• To develop and evaluate a predictive model for imminent stop sign violations.
• To assess the effectiveness of different stop sign warning algorithms in detecting violations.
• To compare braking demands against maximum braking capabilities under various conditions.

Main Methods:

• Utilized data from the 100-Car Naturalistic Driving Study and event data recorders from real-world crashes.
• Developed and evaluated three hypothetical stop sign warning algorithms (early, intermediate, delayed) using logistic regression.
• Assessed violation detection accuracy and braking performance based on required deceleration parameter (RDP) and brake application.

Main Results:

• All three algorithms detected violations, with the early algorithm providing the quickest detection but higher false alarms (22.3%).
• The delayed algorithm significantly reduced false alarms to 3.3% while still detecting violations.
• Vehicle stopping capability varied significantly with surface conditions; Automatic Emergency Braking (AEB) showed potential to improve stopping success rates, even in poor conditions.

Conclusions:

• I-ADAS has the potential to integrate stop sign violation detection algorithms effectively.
• Further validation with larger datasets is recommended for developing more comprehensive algorithms.
• The findings highlight the importance of considering braking capabilities and surface conditions in I-ADAS design.