Project R-5470

About the validity of road user behaviour and near accidents as predictors of the number and the nature of road accidents. (Research)

Traditional analysis methods in the domain of traffic safety rely strongly on crashes as measurement units. However, important limitations are related with the use of crash data. The occurrence of crashes is a relatively rare event, which means that registered variations in crash numbers (over time, between locations...) are at least partly due to chance. Furthermore, crashes are not always correctly reported and - most importantly - too little is known about crash-preceding factors and mechanisms. The main objective of this research is to improve the understanding of crash occurrence by defining relations of these crashes with crash-preceding events and higher-order conditions that influence the occurrence of these preceding events. An improved knowledge hereof would contribute to the establishment of conceptual, law-like relationships between variables describing features of the traffic system (roadway, vehicles and human (inter)actions) and the level of traffic safety. My research will consist of four parts, each of them related to particular research questions on (1) the validation of traffic conflicts as safety indicators, (2) error recovery strategies, (3) validating relationships between nearcrash events and more common behaviour and (4) assessing the 'self-explaining' nature of road environments.

01 October 2014 - 30 September 2017