This paper describes a method, the countermeasure-based program, to identify accident prone locations. The objective of the method is to specify locations that can be treated by road improvements irrespective of their total number of accidents. To account for the random variation inherent in accident occurrence, the method utilised the Empirical Bayes technique. The paper describes the application of the technique to signalised junctions' accident data in Canada and comparison with the traditional accident black-spot program: many locations identified by the countermeasure-based approach because they had well-defined accident patterns were not identified in the black-spot approach. It is hoped that the use of this method would facilitate the diagnosis of accident prone locations and the selection of effective countermeasures.
Samenvatting