Analysis of Microstructures in Traffic Jams on Highways Based on Drone Observations

We investigate vehicle trajectories on a three-lane road segment of a highway which were obtain from drone observations. Based on these comprehensive datasets, we study distance headways to preceding vehicles and develop a method to quantify the local traffic density on a short distance ahead of a vehicle. We leverage these metrics to study a local traffic jam. The new method indicates increased traffic density ahead of a vehicle about 100 meters or 15 seconds before a vehicle is affected by the traffic jam. We also study the connection between vehicle speed and distance headway or traffic density, respectively, and quantify them by conditional probabilities and conditional means. On the one hand, the results give insights into microstructures of traffic jams. On the other hand, the novel method for local density calculation may be applied in vehicles to warn drivers of upcoming high density traffic situations which improves driving safety.