Modeling the impact of various variables on severity of crashes involving traffic barriers

摘要

Traffic barriers are in widespread use all around the world as measures for decreasing the severity of roadside crashes. Considering a comprehensive set of more than 160 variables, this article presents the investigation of the variables affecting the severity of crashes involving traffic barriers. An ordinal logistic regression approach was applied to 10-year crash data collected in Wyoming. The crash data included a wide range of various parameters related to traffic, drivers, and road geometry conditions. The results show that the type of vehicles and traffic barriers has the highest effect on the severity of crashes involved with traffic barriers. The rigid barrier might be the best barrier type for reducing trucks' crash severity, whereas guardrail barrier did not provide the same level of protection for trucks on high truck traffic-volume roads. The presented analytical model is more practical for rural states similar to the Rocky Mountains Region due to its similar conditions. The conclusions of the study will help transportation agencies by (1) guiding the designers when selecting the proper type of traffic barriers based on different road conditions and (2) ranking the risk of various traffic barrier segments for prioritizing the most hazardous ones for future improvements.