Abstract

Pedestrians are among the most vulnerable victims of road traffic accidents. Establishing an unsignalized"br" pedestrian crossing at intersections occasionally results in a high crash risk due to the fact that many vehicle"br" drivers do not heed the legitimate right of way of pedestrians, either deliberately or because of some kind of"br" distraction, speeding or deficiencies in the traffic environment. The primary objective of the OBSERVE project"br" was to develop a novel approach for evaluating crosswalks based on data from observed pedestrian-vehicle"br" driver interactions and local site conditions. Within the project, 85 unsignalized pedestrian crossings in the cities"br" of Graz and Vienna were investigated by means of video observation. The trajectories of different road user"br" categories were analysed to obtain information on driving and walking speeds, traffic behaviour, time gaps etc."br" That information was subsequently used to model driving behaviour. For the modelling process, data from 54"br" zebra crossings were used. A beta-regression model identified the parameters ‘pedestrian crossing type’ and"br" ‘pedestrian crossing width’ having the highest influence on the stopping probability.