In this project we developed a taxonomy of human errors during driving for the Bundesanstalt für Straßenwesen (BASt, German Federal Highway Research Institute). Following a systemic approach, we recognized accidents as a consequence of system components and process interactions, namely the interaction of vehicle, driver, driving task, and environment. As a basis for such classification, accidents were not only analyzed in terms of erroneous results, but also in terms of the dangerous behavior which could lead to accidents. Similarly, behavior which corresponds with normal, accident-free driving was also considered.
In the project, we combined a reappraisal of the current state of driver error classification research and theory construction with our own empirical studies. To test the model assumptions of the rendered taxonomy of errors, we re-analyzed data from studies of naturalistic driving (e.g. 100CarStudy and PROLOGUE) and conducted a validation study in our simulator. Naturalistic driving studies investigate normal driving in real conditions and thus provide a diversification of the usual analytical basis, i.e. accident reports, to observe accident-free driving behavior and situations with events that have not yet led directly to an accident. With the simulator study, the sub-elements of the taxonomy were systematically controlled and investigated. The project results support subsequent conclusions regarding measures of error avoidance, particularly the use of driver assistance systems.
Competencies |
Services |
|
