Tools for Naturalistic VRU Study - Hands-on Manual

Accident data from official records such as police reports or emergency room data is affected by a large degree of underreporting, particularly for vulnerable road users. Naturalistic data collected automatically from the road users can be a means to overcome this problem. The large amount of collected data, however, necessitates the use of technology to facilitate data processing and analysis. This manual describes the three smartphone apps developed with the aim of getting more insight and knowledge in accidents among vulnerable road users based on naturalistic data. The apps and source codes are available via https://bitbucket.org/aauvap/vrumonitorapp/.

The VRUMonitor app was developed to detect accidents based on an analysis of the motion data of the phone. The underlying idea is that abnormal motion occurs in the event of an accident and that this motion differs from everyday activities. Due to the fact that different smartphones contain different motion sensors it is a complex task to make a fully functional system across all smartphone brands and models. Therefore, it has not yet been tested on a larger scale. In addition, the SafeVRU app was developed for self-reporting of accidents and near-accidents and has been used by more than 400 participants.