Using naturalistic data to study time headway behind motorcycles and other vehicles

Winkelbauer, M; Donabauer, M.; Pommer, A.; Jansen, R.

Roughly 10% of motorcycle accidents in Austria are rear-end collisions. One explanatory factor for this could be short time headway, hence this research project. Naturalistic research provides a suitable basis for studying the distance to various kinds of lead vehicles. Our study utilizes the dataset created in the first large-scale naturalistic driving study in Europe (UDRIVE) and combines this with site-based speed data from Austria.

Data on cars was filtered from the UDRIVE dataset and analyzed for the presence of a lead vehicle, with the pertinent information provided by ‘Mobileye’, a real-time video processing unit. The analysis included 1242 h of driving recorded at a frequency of 10 Hz. The difference observed between distances behind cars (1.1 s) and motorcycles (1.2 s) proved to be small.

Almost twenty million records from roadside radar speed recorders in Austria collected in 2015 and 2016 were then compared with this data. In general, the two methodologies delivered similar results: time headway decreases with increasing driving speed in both datasets. The analysis of the UDRIVE data also showed minor differences in time headway among drivers in the six UDRIVE countries (France, Germany, Poland, Spain, The Netherlands, United Kingdom). Both datasets suggest that time headway is shortest behind trucks. However, the hypothesis that car drivers maintain less time headway behind motorcycles is not supported by either the UDRIVE or the site-based data.

In this paper, we will discuss both these results as well as the strengths and weaknesses of both data collection methods.

Published in
Safety Science

SWOV publication

This is a publication by SWOV, or that SWOV has contributed to.