A reduction in accidents, particularly secondary ones, is often cited as one of the primary benefits of Freeway Traffic Management Systems (FTMS's) such as that implemented on a segment of Highway 401 in Toronto, Canada. Safety evaluation of such systems is often confounded by the changes in traffic intensity resulting from the FTMS and other causes. Thus, the conventional before and after method is not always appropriate for such evaluation studies. The fundamental objective of the research was to evaluate the safety effectiveness of that system using microscopic accident potential models, a sophistication made necessary in this case because the implementation of the FTMS coincided with a fairly sharp decrease in traffic volumes which, by itself, could have caused the significant decrease observed in the accidents of interest. For the research summarized in this paper, the safety effect of the FTMS was estimated from the parameters of regression models calibrated for the pre-implementation period (1989-1990) and the post-implementation period (1991-1992). In view of the relative novelty of methodological aspects such as the use of generalized linear modelling with a negative binomial error structure and the use of hourly volumes (as opposed to AADT's) as a measure of traffic intensity, a secondary, but important, objective was to use the research to demonstrate the development and application of these models. Separate models were calibrated for express, collector and regular divided roadway sections, and for rear-end, non-rear-end and all injury accidents. For rear-end injury accidents, model parameters were estimated for day and night conditions. The aggregate results show a substantial decrease in the expected number of rear-end injury accidents on all three types of roadway following the implementation of the FTMS. It is expected that the results would aid in the planning, design and operation of future systems and in refining the operation of existing ones. (A)
Samenvatting