Assessing Transferability of Highway Safety Manual Crash Prediction Models to Data from Italy

For decades, crashes have been studied as discrete events with the focus on the circumstances of the crash. This type of analysis has been used to identify the characteristics of roadway features associated with higher crash experience, but other factors, such as traffic volumes, driver characteristics, land use, and environmental conditions, are also needed to explain or describe crash events. The Highway Safety Manual (HSM) provides a predictive method to estimate the expected average crash frequency of a site in given geometric and geographic conditions over a specific period for a specific annual average daily traffic volume. The study presented here investigated whether the modeling results closely matched the crash records. The HSM algorithms were used to assess transferability as a whole. The results suggested that implementing the HSM techniques should foster the development of local safety performance functions and accident modification factors. Calibration preserved the original HSM model form and the relationship between independent variables and crashes. To adjust the base predicted crash frequency to meet the current conditions, the accident modification factor calculations for lane width, horizontal curves, and vertical grades were made. Crash types (head-on and side collisions, single-vehicle crashes, and rear-end collisions) were investigated on the basis of the vertical grade and the curvature indicator. The estimated model provides planners and designers with a tool better able to target and select countermeasures to address these specific aspects and results in improved project selection and improved safety.
Transportation Research Record: Journal of the Transportation Research Board
Francesca Russo, Mariarosaria Busiello, Salvatore A. Biancardo, and Gianluca Dell’Acqua
Anno del prodotto: 
Volume e Anno: 
Volume 2433
Pagina inizio: 
Pagina Fine: 
Societa' Italiana Infrastrutture Viarie - S.I.I.V. - Cod.Fis. 93024730421