The purpose of this research investigation was to analyse the effects of environment, topography, geometry, and highway and rail traffic patterns with respect to rail-highway grade cross sings accidents in rural areas. The mathematical tools of factor analysis and regression analysis were used to develop models for predicting the relative hazard at a railway grade crossing. These models are based on rail volume, highway volume and roadside distractions, such as houses, business and advertising signs. To evaluate the proposed mathematical relationships, it was necessary to collect sufficient data on many variables deemed to have an influence on safety. Therefore 56 variables were measured at the 289 accident locations and 23 variables at the 241 non-accident locations. Previous research efforts were concerned either with long period accident experience or with before-and-after studies of the various protection devices. In this research, locations which experienced accidents in a two-year period were compared to non-accident locations. The results of this study can be used to determine the type of protection which a crossing warrants.
Abstract