Collision Models for Multilane Highway Segments Incorporating the Effects of Curbs

Abstract

The main objective of this study was to develop valid statistical collision models for multilane highway segments with or without curbs. For this, road geometric data, traffic data, and collision data for the three years were collected. The data include 2,274 collisions and 885 injury collisions that occurred on 191.85 miles of 199 directional segments.

A new modeling method of introducing variables into the model one by one in a multiplicative form was applied. A nonlinear optimizing algorithm for estimating parameters using a negative binomial log likelihood function was adopted for the modeling. The functional form of the variable to be introduced was determined based on the relationship between the recorded number of collisions and the number of collisions predicted by the current model without the variable. The integrate-differentiate method was applied to find candidate functional forms for each variable. Model selections were based on the -2 log likelihood and BIC statistics. The cumulative residuals (CURE) plot method was adopted for checking the goodness of fit of the models.

As a result of the modeling efforts, the annual average daily traffic, access point density, shoulder width, and shoulder type variables were introduced to the final model for total collisions. The same variables except the shoulder type variable were introduced to the injury collision model. Overall, then, it appears that curbs mean fewer total collisions and no change in injury collisions as compared to no curbs on the sampled road segments. The models developed in this study were based only on the data for North Carolina and limited number of variables. The developed models can be improved in the future by collecting data on more miles, by bringing more explanatory variables into models, and by using the data from other states.

Additionally, the characteristics of vehicles speeds on multilane highways were analyzed and compared. The results showed that the mean speeds for the non-curbed sites were about 2 to 3 mph higher than those for the curbed sites.