Towards a pro-active model for identifying motorway traffic risks using individual vehicle data from double loop detectors

This paper presents a methodology to develop motorway traffic risk identification models using individual vehicle traffic data, meteorological data and crash database for a study site at a two-lane-per-direction section on motorway A1 in Switzerland. We define traffic situations (TSs) representing traffic status for three-minute interval and traffic regimes obtained by clustering TSs. The models are traffic regimes - based and are developed using Regression Trees to identify rear-end collision risks. Interpreting results shows that speed variance on the right lane and speed difference between two lanes are the two main causes of rear-end crashes. We also compare the results obtained from three-minute TSs with the results obtained from five-minute TSs using the same methodology.