Abstract:
The increasing need for mobility of people leads to steady increase of spatiotemporal congestion phenomena in more and more congested road networks when using vehicles. New vehicle technologies and developments such as electromobility and autonomous driving must deal with and master this reality on the roads. A deeper understanding of all congestion phenomena, their causes, origin and spatiotemporal development therefore provides an essential scientific basis. For comprehensive and microscopic traffic detection a variety of technologies have been developed and used to this day. In this work, access to large amounts of traffic data from different data sources (floating car data, drone data and detector data) distributed across Europe and the U.S. enables reliable and detailed investigations of various congestion phenomena on highways. As a central result, the empirical evidence is provided for the existence and characteristics of spatially bounded and temporally limited traffic disruptions before traffic breakdowns. These traffic disruptions occur in free traffic flow at highway bottlenecks and have been proven based on floating car data. They exhibit general characteristics and common features and can exist as spatiotemporally propagating traffic structures. Hereby, missing evidence is provided for the concepts of traffic flow formulated in traffic theories. Since floating car data are not lane-specific resolved today, highly accurate and complete drone data are used to prove the occurrence of lane-dependent traffic breakdowns. Congested traffic structures and traffic structures with high traffic density are observed on all three lanes at different points in time and locations. For this purpose, a traffic density method based on moving averages and the spatial distances between consecutive vehicles is revealed. Moreover, detector data are used to prove the traffic-phase- and lane-dependency of time gaps between vehicles. The probability of merging and lane-changing opportunities on highways also exhibit traffic-phase- and lane-dependent characteristics. The traffic phase of synchronized flow turns out to be the most unsuitable traffic phase for finding a sufficiently large time gap for an (automated) vehicle to merge onto highways or change lanes. In addition, a waiting time model is developed to calculate the waiting time that must pass at a spatial location on the highway to find a sufficiently large time gap to merge onto the highway or change the lane. The waiting time exhibits traffic-phase-dependent characteristics. The results of this work make an important contribution to the better understanding of traffic breakdowns in real traffic and to the improvement of traffic reconstruction. They can serve as requirements for traffic theories and models and can be used in automated vehicles, driver assistance and navigation systems.