Survival Analysis-Based Modeling of Urban Traffic Incident Duration: Shanghai Case Study, China
Because of the intensively use of urban expressway for a long time, any small disturbance like traffic incidents may cause a large-scale congestion on expressway network in Chinese metropolises. How to know well about the formation mechanism and propagation characteristics is being increasingly perceived as the most challenging problems for safety expressway management. Traffic incident duration is one of most important parameters to reflect traffic congestion intensity, and numerous measures have been developed to describe the characteristics of traffic incidents and the vast majority of these studies use data mining. Though these measures describe what objective elements of the environment may influence traffic incident, the how question-via what way these physical features affect traffic incident-is largely unexplored. Based on the historical traffic data of viaduct expressways in Shanghai, this paper introduces survival analysis into the analysis of traffic incident mechanism, and a survival Analysis-Based Modeling of urban traffic incident duration is presented. This model first analyzes the time attributes of many traffic incident samples, and employs nonparametric regression based Kaplan-Meyer model to estimate hazard-based traffic incident duration time. Then, the key influence factors of traffic incident are divided into five types and the spatial-temporal distribution characteristics of traffic incident duration time are analyzed. Finally, COX regression is used to model the co-evolution between multidimensional influencing factors of traffic incident duration. The key characteristic parameters of expressway incident management in Shanghai are optimized to analyze the evolution mechanism of incident duration. The result shows that, for different type of influencing factors, the spatial-time distribution of traffic incident duration in Shanghai expressway exists significant difference, and factors like day & night, incident type, related vehicle number, related lane number, location, bottleneck and trailer will affect the incident duration significantly.
-
Supplemental Notes:
- This paper was sponsored by TRB committee ABE90 Transportation in the Developing Countries.
-
Corporate Authors:
500 Fifth Street, NW
Washington, DC United States 20001 -
Authors:
- Zhang, Lun
- Shi, Yicheng
- Yang, Wenchen
- Liu, Pei
- Rao, Qian
-
Conference:
- Transportation Research Board 93rd Annual Meeting
- Location: Washington DC
- Date: 2014-1-12 to 2014-1-16
- Date: 2014
Language
- English
Media Info
- Media Type: Digital/other
- Features: Figures; Maps; References; Tables;
- Pagination: 18p
- Monograph Title: TRB 93rd Annual Meeting Compendium of Papers
Subject/Index Terms
- TRT Terms: Bottlenecks; Case studies; Crash characteristics; Data mining; Expressways; Incident management; Mathematical models; Periods of the day; Traffic congestion; Traffic incidents; Urban areas
- Geographic Terms: Shanghai (China)
- Subject Areas: Data and Information Technology; Highways; Operations and Traffic Management; I73: Traffic Control; I81: Accident Statistics;
Filing Info
- Accession Number: 01520350
- Record Type: Publication
- Report/Paper Numbers: 14-4393
- Files: TRIS, TRB, ATRI
- Created Date: Mar 27 2014 3:38PM