Parts | Days | Selection | Search | Updates | Downloads | Help
SOE: Fachverband Physik sozio-ökonomischer Systeme
SOE 12: Urban Systems and Traffic Flow
SOE 12.3: Talk
Wednesday, March 20, 2024, 10:30–10:45, MA 001
Fast & Furious: Extreme events and non-Gaussian velocities in urban car traffic — Moritz Piepel1, •Malte Schröder1, Angelika Hirrle2, and Marc Timme1,3 — 1Chair for Network Dynamics, Institue for Theoretical Physics and Center for Advancing Electronics, TUD Dresden University of Technology — 2Chair of Traffic Process Automation, Institute of Traffic Telematics, TUD Dresden University of Technology — 3Lakeside Labs, Klagenfurt
The majority of traffic flow models and observations focus on traffic dynamics on highways, assuming Gaussian velocity distributions typically observed in time-aggregated measurements. Here, we analyze individual velocity measurements in urban car traffic based on 145 induction loop detectors throughout the city of Dresden, observing over 340 million vehicle velocities in total. We find that velocity distributions in urban traffic are non-Gaussian, with frequent extreme velocities, independent of the local speed limit. Gaussian distributions significantly underestimate the frequency of extreme velocities and their consequences. For example, the number of speeding violations observed in the data would be valued in fines of 800 million Euros annually, more than 300 times the actual fines collected in Dresden. These observations seem to be generic and are confirmed by data on speeding violations in the city of Cologne. Our findings shed a new light on urban traffic modeling and may have implications for road safety regulations, the design of road infrastructure, and speed limits to ensure safe urban mobility.
Keywords: Urban mobility; Traffic flow; Extreme events; Non-Gaussian distributions