摘要: In 2019 the European Parliament updated Directive 2008/96/EC on Road Safety Management. In the updated Directive, they introduced the concept of Network-wide Road Safety Assessment (NWRSA), which is a network screening that must be based on both visual inspections of the road and accident analysis. The procedure should provide a classification of the different road sections analyzed into at least three different risk levels (low, medium, and high). This paper investigated eight safety analysis procedures that are eligible to be included in the NWRSA process to identify elements in common and in conflict among them. These procedures have been applied to a two-lane two-ways rural highway stretch of about 10 km. Due to the differences in the parameters, algorithms, and processes considered in each procedure, the results of each evaluation may differ, both considering the typology of the output, both considering the measure of risk for a specific segment. To provide the road authorities with effective assessment tools for all road segments inside their network as a function of data available, the results of the procedures are analyzed and discussed both qualitatively and quantitatively using statistical analysis such as Kendall W and non-parametric Kruskal-Willis, to test for concordance and significative difference between the analyzed procedure respectively.
Moreover, the main characteristics of the procedures are highlighted and discussed, such as the required data and the considered segmentation.
The results show that some procedures lead to results that do not differ significantly. Therefore, they can be used as alternative procedures to assess different road stretches according to specific data availability, while others not.