Point-process modeling of secondary crashes.

Area of Science:

• Traffic safety analysis
• Statistical modeling of transportation incidents

Background:

• Secondary crashes, occurring after primary incidents, pose extreme dangers to responders and victims.
• Understanding crash dynamics is crucial for improving highway safety and incident response.

Purpose of the Study:

• To develop and apply a self-exciting temporal point process model to differentiate primary and secondary highway crashes.
• To estimate the prevalence of secondary crashes and associated queue times using real-world data.

Main Methods:

• Utilized a self-exciting temporal point process model, distinguishing secondary crashes with a self-exciting function and primary crashes with a background rate.
• Modeled the background rate using stationary, sinusoidal non-stationary, and piecewise non-stationary functions to capture traffic variations.
• Fit the model to Florida Department of Transportation crash data from Interstate-4 (I-4) highway (2015-2017).

Main Results:

• The sinusoidal non-stationary background rate function demonstrated superior fit to traffic data, accurately reflecting daily and weekly crash peaks.
• Secondary crashes were found to constitute up to 15.09% of traffic incidents on the I-4 highway.
• Model parameters enabled estimation of secondary crash probabilities and queue times.

Conclusions:

• A self-exciting temporal point process model effectively distinguishes secondary crashes from primary ones.
• Sinusoidal modeling of background rates accurately captures traffic-related crash patterns.
• Secondary crashes represent a significant portion of highway incidents, highlighting the need for targeted safety interventions.