Prediction of urban road safety

The relationship between accident counts and environmental influences was quantified by analyzing the main road network of Berlin and subsequently calculating generalized regression models. This analysis was carried out individually for road sections and intersections, while further differentiating intersections by type of traffic regulation. In order to assess the transferability of the quantified functional relationships, these were validated on the basis of a subset of the main road network of Dortmund. The regression models fitted for Berlin and Dortmund allow an estimation of accident counts as a function of numerous operational and design-related variables. Thus, for example, the impact of short- and long-term changes of traffic volume on traffic safety can be depicted. Moreover, the developed models may be integrated into safety concepts, in order to derive rankings by priority as well as countermeasures and thereby effectively support road safety efforts. Language: de