Safety Performance Assessment of Freeway Interchanges, Ramps, and Ramp Terminals

The paper describes the development of new safety performance functions for interchanges, ramps and ramp terminals for Ontario freeways, using negative binomial regression modelling that relates collision frequency to traffic volumes and basic entity characteristics. Also presented is the application of these safety performance functions for network screening using two varieties of the potential for safety improvement (PSI) index method, one based on expected collision frequency and one based on expected collision frequency in excess of what is considered normal. The rankings of screened sites based on the two methods are compared. A third method, which is based on an index of a high proportion of a specific collision type, is applied to ramp terminals by way of illustration to identify those sites with high proportions of specific collision types. This method does not require safety performance functions or traffic volumes but does require the application of some fairly intricate statistical methodology. A comparison of the rankings so obtained with those derived by applying the PSI methods for a specific collision type suggests that the method of screening for high proportions of specific collisions can be a useful alternative to the PSI index method where safety performance functions and/or traffic volumes are not available since, unlike the PSI Index method, it does not require these inputs in doing network screening for specific collision types. This work has been developed as part of the continuing efforts of the Ministry of Transportation Ontario to establish a state-of-the-art knowledge base on the interaction of human factors, highway design and highway safety that can be put into practical use. The body of knowledge that is being assembled is referred to as the “The Science of Highway Safety”.