Heterogeneity of Microscopic Congested Traffic Data Based on Drone Measurements

Abstract

We study vehicle trajectories at the onset and existence of traffic congestion and reveal its microscopic features on separate highway lanes. Drone observations of microscopic data of moving vehicles have been made available on three-lane road segments of German highways. Based on these detailed empirical traffic data we reveal heterogeneity and complexity of congested traffic and discuss its consequences. E.g., to perform a safe and comfortable driving behavior by driver assistance systems or automated vehicles lane-level traffic states should be adapted. A congested and dense traffic state only on the left lane of a three-lane highway could be a serious danger. Moreover, we propose an empirical method to calculate local traffic densities that could be used to warn vehicles in advance about high preceding densities. We leverage that concept to study a local traffic jam and discuss its lane-level properties. We reveal the heterogeneity of high local density structures on separate highway lanes.