Modeling crash frequency of heavy vehicles in rural freeways

In the recent years, Prediction models of crash frequency have become a criterionfor accident rate and are remarkably considered in safety improvement plan.Heavy vehicles are one of the main causes of accidents in the world as 1 of the 9accidents is involving them. In this research, a prediction model was developedby using heavy vehicles accident data collected from Karaj-Qazvin freeway. Inthe model, important factors of occurrence of accident including climaticcondition, road surface condition, natural light condition, average vehicle speed,average daily traffic and 5 dangerous violations commitment were considered.Poison regression and negative binomial regression models were applied formodeling purposes. After evaluation of models, it concluded that factors such asnatural light condition, heavy vehicles daily volume, speed limit violation andvehicle defects especially in lighting system during darkness had great impact onheavy vehicles accident occurrence.

Keywords:

heavy vehicles accident frequency , poison regression model , negative binomial regression model