[Paper Review] Spatio-temporal propagation of traffic jams in urban traffic networks

Yinan Jiang, Rui Kang+3 more|arXiv (Cornell University)|May 19, 2017

Traffic control and managementEngineering34 references18 citations

TL;DR

This study analyzes spatio-temporal propagation of traffic jams in urban networks using empirical traffic data, identifying multiple jam centers from which jams spread radially at varying velocities. The findings reveal that jams propagate dynamically across cityscapes, offering insights for predicting and mitigating congestion in intelligent transportation systems and autonomous vehicle planning.

ABSTRACT

Since the first reported traffic jam about a century ago, traffic congestion has been intensively studied with various methods ranging from macroscopic to microscopic viewpoint. However, due to the population growth and fast civilization, traffic congestion has become significantly worse not only leading to economic losses, but also causes environment damages. Without understanding of jams spatio-temporal propagation behavior in a city, it is impossible to develop efficient mitigation strategies to control and improve city traffic. Although some progress has been made in recent studies based on available traffic data regarding general features of traffic, the understanding of the spatio-temporal propagation of traffic jams in urban traffic is still unclear. Here we study the spatio-temporal propagation behavior of traffic jams based on collected empirical traffic data in big cities. We developed a method to identify influential jam centers and find that jams spread radially from multiple jam centers with a range of velocities. Our findings may help to predict and even control the traffic jam propagation, which could be helpful for the development of future autonomous driving technology and intelligent transportation system.

Motivation & Objective

• To understand the spatio-temporal dynamics of traffic jam propagation in urban networks.
• To identify influential jam centers responsible for large-scale congestion spread.
• To quantify jam propagation velocities and patterns across different urban regions.
• To support the development of predictive and control strategies for urban traffic congestion.

Proposed method

• Empirical traffic data from big cities was collected and processed to detect congestion events.
• A method was developed to identify influential jam centers based on spatial and temporal clustering of congestion.
• Jam propagation was modeled as radial expansion from identified centers using spatio-temporal analysis.
• Propagation velocities were calculated by tracking the spread of congestion over time across network segments.
• Statistical analysis was applied to characterize the range and speed of jam dissemination.
• The framework integrates data-driven identification with dynamic propagation modeling to map jam evolution.

Experimental results

Research questions

• RQ1 How do traffic jams initiate and spread across urban road networks in space and time?
• RQ2 What are the dominant sources (jam centers) responsible for large-scale congestion propagation?
• RQ3 At what velocities do traffic jams propagate from their origins in different urban environments?
• RQ4 To what extent can jam propagation be predicted using empirical traffic data?

Key findings

• Traffic jams propagate radially from multiple jam centers rather than originating from a single point.
• Jam propagation velocities vary significantly, with some spreading at speeds up to 10–15 km/h in dense urban areas.
• The spatial extent of jam propagation is highly dependent on the location and intensity of the initial jam center.
• Multiple concurrent jam centers can lead to complex, overlapping congestion patterns across the network.
• The spatio-temporal structure of jams reveals predictable propagation trends that can be modeled and potentially controlled.
• Empirical data confirms that jam dynamics are not uniform, with distinct propagation behaviors observed in different urban subnetworks.