BACKGROUND

To reduce the effects of a traffic accident it is essential to transport the casualties to the right care facility with minimal delay. For (single) accidents involving Vulnerable Road Users (VRUs) it is often a problem to know about and report the accident and its location swiftly, and thereby the rescue gets delayed. EU's eCall regulation requires future cars to be equipped with Automatic Crash Notification (ACN), which notifies 112 with the vehicle’s exact location (GPS) in case of a crash.

Today’s smartphones are equipped with high quality movement sensors (accelerometers, gyroscopes, magnetometers), GPS and SMS functionality, i.e. key components of an ACN. Thus, the aim of this study was to explore the feasibility to develop ACN for VRUs based on regular smartphones. 

METHODSWe have focused on recording normal activity from bicyclists, horse-back riders, and ATV (All Terrain Vehicle) users, i.e. three different VRU categories who could benefit from an ACN system. Based on these recordings and simplified crash tests/simulations we have developed algorithms for incident detection, i.e. activities which are recognized as non-normal and should be reported as a potential accident.

RESULT A fully functional ACN app for bicyclists (available from Google Play) and a beta prototype for horse-back riders have been developed. ATV use may be the most challenging of the three VRU groups, as ATVs are often used at low velocities and may rollover without causing high impact to the driver. A first prototype of an ACN algorithm for ATV users resulted in 13 false alarms during 49 hours, i.e. about one every fourth hour. The current development is focused on improving the accuracy of incident detection for ATVs, firstly by decreasing the number of false alarms.

CONCLUSIONSThis work demonstrates the feasibility of smartphone based ACN systems for VRUs.