Automatic Near-Stationary Traffic State Identification Based on PELT Changepoint Detection

The existence of stationary traffic states has been widely used as a key assumption in studying traffic flow models and analyzing transportation network problems. However, many empirical studies do not properly use near-stationary states to establish or validate their models. In addition, there lacks a systematic method to identify near-stationary states in the transportation literature. This paper starts with a comprehensive introduction of stationary states and presents an automatic near-stationary state identification method. In this method, the authors first apply a pruned exact linear time (PELT) changepoint detection algorithm to split flow and occupancy sequences into multiple candidate intervals, in each of which averages of flows/occupancies stay relatively constant over time. After deriving characteristics of the detected candidate intervals, the authors present two identification criteria based on Cassidy’s method to extract near-stationary states from the candidates. The authors provide two ways of representing near-stationary states and apply near-stationary data on the fundamental diagram calibration. The results demonstrate that the identified near-stationary states can yield a nearly scatter-free flow-occupancy relation, and further lead to a well-defined triangular fundamental diagram.

• Supplemental Notes:
  • This paper was sponsored by TRB committee AHB45 Standing Committee on Traffic Flow Theory and Characteristics. Alternate title: Automatic Near-Stationary Traffic State Identification Pased on PELT Changepoint Detection
• Corporate Authors:

  Transportation Research Board

  500 Fifth Street, NW
  Washington, DC  United States  20001
• Authors:
  • Yan, Qinglong
  • Sun, Zhe
  • Gan, Qijian
  • Jin, Wenlong
  • 0000-0002-5413-8377
• Conference:
  • Transportation Research Board 95th Annual Meeting
  • Location: Washington DC, United States
  • Date: 2016-1-10 to 2016-1-14
• Date: 2016

Language

• English

Media Info

• Media Type: Digital/other
• Features: Figures; References; Tables;
• Pagination: 20p
• Monograph Title: TRB 95th Annual Meeting Compendium of Papers

Subject/Index Terms

• TRT Terms: Algorithms; Methodology; Traffic density; Traffic flow; Traffic speed
• Uncontrolled Terms: Fundamental diagram
• Subject Areas: Data and Information Technology; Highways; Operations and Traffic Management;

Filing Info

• Accession Number: 01592758
• Record Type: Publication
• Report/Paper Numbers: 16-2847
• Files: TRIS, TRB, ATRI
• Created Date: Mar 4 2016 5:05PM