Macroscopic Traffic Dynamics in Urban Networks during Incidents

Amini, Sasan; Tilg, Gabriel; Busch, Fritz · 2020 · Crossref

DOI: 10.36227/techrxiv.12865988.v1

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the challenge of characterizing macroscopic traffic dynamics in urban networks during non-recurring congestion, specifically incidents. While the Macroscopic Fundamental Diagram (MFD) is a established tool for understanding network performance under recurring conditions, its applicability during incidents is limited by the emergence of hysteresis loops and multivaluedness. The authors aim to bridge this gap by proposing a framework that links the shape of the MFD during incidents to the criticality of the affected network links. This approach seeks to enable incident detection and traffic management strategies even under limited sensor coverage, moving beyond methods that rely on extensive historical data or dense sensor networks. To investigate this relationship, the authors introduce a Criticality Score (CS) for network links, defined by the reduction in the total number of shortest paths between all node pairs when a specific link is closed. This metric serves as a proxy for network redundancy. The study employs a microscopic traffic simulation using SUMO on a 5x5 unidirectional grid network with 60 links and signalized intersections. The experimental design includes a base scenario representing dynamic user equilibrium and seven incident scenarios where specific links are closed. Drivers are assumed to learn of closures only upon reaching the affected area, simulating unexpected incidents. The authors quantify the resulting MFD characteristics using variables such as capacity loss, loop height, loop width, loop area, and average speed within the hysteresis loop. The results confirm the authors' postulation that links with higher Criticality Scores impose larger hysteresis loops and greater capacity losses on the MFD. Specifically, the closure of links with similar CS values resulted in comparable deviations in the MFD shape and similar patterns in trip completion rates. The study observed that higher CS values correspond to lower network redundancy, leading to uneven congestion distribution and significant spillback. Conversely, networks with higher redundancy exhibited more stable MFDs. Additionally, the authors identified a consistent bifurcation point across all scenarios, where the incident MFD diverged from the base MFD at an occupancy of approximately 8% and a flow of 400 veh/h/ln, suggesting a network-specific instability threshold. The significance of this work lies in the development of a framework for incident detection in urban networks using limited sensor data. By clustering links based on their criticality and monitoring MFD deviations, traffic operators can detect incidents on unsensed links by observing traffic dynamics elsewhere in the network. The findings imply that network redundancy is a key factor in mitigating the impact of incidents on macroscopic performance. However, the authors note that the results depend on assumptions regarding network topology, origin-destination patterns, and driver route choice behavior. Future research is required to validate these findings in more complex network structures and with realistic driver adaptation strategies, such as real-time rerouting information.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• traffic density