Developing behavioral models for driver gap acceptance at priority intersections

Abstract

Priority intersections form a major element in road networks since significant portions of traffic accidents and delays occur at them. Nevertheless, there is lack in studies directed to evaluate the performance of these junctions and to understand driver behavior at them. Previous research has acknowledged that priority junctions can be studied within the context of driver gap acceptance behavior and has raised the importance of incorporating driver, vehicle, and trip attributes in gap acceptance studies related to these junctions. However, limited effort has been directed to this subject. The quantitative effects of driver, vehicle, and trip attributes were not evaluated and the reported results in literature about the effects of other typical traffic attributes aer inconsistent.

In this study Binomial Logit Behavioral Models are developed for diver gap acceptance at priority intersections. The study has investigated the effect of the main driver, vehicle, trip, gap, and traffic attributes on driver gap acceptance behavior at T-intersections. Left turns form major road and right and left turns from minor road are considered. Data needed to calibrate the models were collected using field administered questionnaires and video cameras. More than thirty models are reported and discussed in the study. Along with the basic traffic and delay attributes such as the gap size, the speed of the approaching vehicle, the total delay imposed on the driver at minor road, and the traffic volume at minor road the driver, trip and vehicle attributes such as driver age, sex and accident experience, trip duration, and vehicle occupancy were found to be significant factors in explaining driver gap acceptance behavior. The study has also concluded that binary choice models other than the binomial Logit could provide an adequate framework for modeling driver gap acceptance process.

The results of gap acceptance analysis are applicable to different traffic engineering fields. As an example, the study has demonstrated the applicability of the calibrated models to identify sight distance requirements at priority intersections.