This study used the method detailed in Traffic Conflict Techniques for Safety and Operations (US Department of Transportation, 1988) for the identification of average and abnormally high values of traffic conflicts, for 40 three-leg, unsignalised intersections in Amman, Jordan. The negative binomial distribution was used to rank intersections in terms of their hazard potential. This method has been used successfully by various researchers, but they have only considered conflict counts at intersections, and have disregarded the geometric characteristics of the intersections and entering traffic volumes. In this paper the authors investigated the differences in the ranking order of the selected intersections when: (a) intersections with similar or different geometric characteristics (in terms of being divided and/or undivided by kerbed islands on the major approaches) were considered in the analysis and (b) entering traffic volumes were taken into account. The authors found that for any particular conflict type the order of ranking intersections in terms of their abnormal conflict counts, was influenced by the intersections being divided or undivided with kerbed islands. They also found that the incorporation of traffic volume influenced the ranking. As a result, when the traffic conflict technique is used to rank intersections in terms of their hazard potential, it is recommended that the geometric characteristics of the intersections and the traffic volume should be taken into consideration. (a).
Samenvatting