Development of a Crash Prediction Model for Rural Roads in NSW, Australia

An important part of road traffic safety management is the collection, use and evaluation of comprehensive and accurate road crash data. The reasons include: (1) analysis of this data provides better understanding of the operational problems and provides the basis for prediction models for road crash savings; (2) it provides road authorities with sound information to develop road safety strategies and action plans; (3) it is a pre-requisite for the accurate identification of road traffic accident related problems; (4) it is an input into road safety engineering assessment - an important tool to assess the safety status of the entire road network; (5) it is an input into the development of remedial measures; and (6) it is necessary to evaluate the success or effectiveness of safety management plans and actions. By studying accident statistics and patterns and their relationship to road attributes, road managers can concentrate on the improvement of those features that tend to reduce the safety amenity of a road. Numerous research projects have established relationships between crashes and geometric road elements, operating speed and traffic volumes. These studies have led to the development of crash prediction models or operating speed models, which greatly increase our understanding of interrelated parameters affecting road safety. This paper provides some detail of the project undertaken for the Roads and Traffic Authority, NSW in Australia and defines the approach and method for the derivation and computation of rural roads crash rates for various stereotypes (roadway cross sections with similar design characteristics) and the means for the provision of an extensive database/spatial database of road stereotypes that are apparent through out the State of NSW.