Taxonomy for Driving Errors

In this project we are developing a taxonomy of driver errors on behalf of the German Federal Highway Research Institute BASt. The project started in July 2011. Within a systemic approach, which understands accidents as a consequence of an interplay of system components and processes, i.e. driver, driving task, vehicle, and environment, the taxonomy will not only be based on accidents as consequences of driver errors but will consider behavior, which led to incidents (not yet accidents), and regular, accident-free driving behavior as well.

In the project we combine a review on the state of the art of driver error classification research and theory construction with own empirical studies. We plan to re-analyze data from naturalistic driving studies, e.g. 100CarStudy and PROLOGUE, and to conduct a simulator study in order to test the assumptions of the developed error model. Naturalistic driving studies investigate regular driving behavior in a natural setting and thus provide a rich basis for analysis, including accident-free driving, incidents, i.e. behavior that nearly led to an accident, as well as accidents. The planned simulator study will enable us to investigate aspects of our taxonomy in a systematic and controlled manner. The project results will support informed decisions on measures for driver error mitigation or prevention, especially in the context of driver assistance systems.