Naturalistische Fahrverhaltensbeobachtungen in Deutschland: Analysemethodik und aktuelle Ergebnisse einer Pilotstudie

Kurzfassung

The assessment of Naturalistic Driving Data (NDS) is a valuable approach to quantify spontaneous driving behavior in real traffic conditions. This is especially true for behaviors that cannot adequately be studied under controlled conditions due to legal or ethical reasons (i.e. distraction, fatigue).The large scale database enables researchers on the one hand to describe normal driving behavior and on the other hand to identify accidents and critical driving events and their prevalence in traffic. BASt and FAT jointly follow the vision of a continuous naturalistic driving study in Germany supplementing established accident data bases and therefore improving knowledge about accident causation. This database could ultimately improve the development and evaluation of vehicle-, infrastructure and behavior based road safety measures. The status of two studies that were designed to build the methodological as well as technical basis for a German NDS will be presented. From the methodological side the calculation of risk measures quantifying the additional risk of critical driving events under certain conditions is a basic feature of NDS (bad weather, phoning, etc.). In order to assess prevalence and risk data, a reasonable unit of evaluation must be defined. Neither whole trips nor one- or more-minute episodes are appropriate for the underlying questions. The segmentation of a trip in well- distinguished driving maneuvers as homogeneous action units in traffic provides both, a good and compact description of a trip itself, and a basic unit to calculate prevalence and risk measures of certain critical driving events. The prevalence of critical driving events within the car-following maneuver for different conditions will be reported. On the technical side, an ongoing pilot study pursues the objective to define and establish the processes of data acquisition, data management and data evaluation providing a basis for large-scale NDS. Especially the data acquisition system (DAS) and the data management will be presented. This includes mainly the design of the DAS, the required data for further evaluations, and regarding the database the process for saving and analysing data. Furthermore, first results of an exemplary evaluation of when and how the participants of the study are distracted during their driving tasks will be shown.