This paper presents the results of a laboratory experiment which sought to analyze rear-end collision risk in various traffic conditions as well asvarying road geometric configurations. The CRISS (Inter-University Research Center for Road Safety) interactive fixed-base driving simulator was used. An 8-km, two-lane rural road incorporating 15 different tangents and curves, was designed and implemented in the driving simulator. Four distinct traffic conditions were simulated and thirty-two drivers drove in four traffic scenarios. The differences in speed and position between the drivers vehicle and other vehicles on the road were collected to calculate two safety indicators which had been suggested by the scientific literature for evaluating rear-end collision risk, based on the time-to-collision (TTC)notion. The maximum values of these indicators were then given for an average traffic-flow condition; whilst pointing out that this condition is the most critical one for rear-end collision risk. In scenarios entailing lower traffic volumes, significant models of TTC-based indicators were foundfor tangents with lengths ranging from 400 to 1000m. The independent significant variables were volume-to-capacity ratio, length of the tangent as well as length of the curve approaching the tangent. In traffic scenarios characterized by higher traffic volumes, no significant models were found for any kind of geometric element. This outcome seems to prove that the influence of the geometric features on the rear-end collision risk tends to become negligible whenever the volume-to-capacity ratio reaches high levels. In these conditions driver behavior is exclusively affected by interactions among vehicles.
Abstract