Skip to main content
Home

Main navigation

  • About
    • About TREC
    • Our Staff
    • Our Researchers
    • Contact Us
  • News
    • Latest News
    • Join Our Mailing List
    • Media Coverage
  • Events
    • Upcoming Events
    • Past Events
  • Research and Data
    • Researchers
    • All Projects
    • Final Reports
    • PORTAL: Portland-Vancouver
    • BikePed Portal: National
  • Our Students
User account menu
  • Log in

Unsupervised Approach to Investigate Urban Traffic Crashes Based on Crash Unit, Crash Severity, and Manner of Collision

Principal Investigator:

Farzin Maniei, University of Texas, Arlington

Co-Investigators:

  • Stephen Mattingly, University of Texas, Arlington

Summary:

Both crash frequency analysis (CFA) and real-time crash prediction models (RTCPMs) typically divide a highway into segments with a constant length for data aggregation. Despite the significant impact of the segment length in several previous traffic safety studies, no standard approach exists for determining a recommended segment length for crash data aggregation. This research investigates the im... Both crash frequency analysis (CFA) and real-time crash prediction models (RTCPMs) typically divide a highway into segments with a constant length for data aggregation. Despite the significant impact of the segment length in several previous traffic safety studies, no standard approach exists for determining a recommended segment length for crash data aggregation. This research investigates the impact of fragment size (segment length) on traffic crash data aggregation and establishes a methodology for determining a recommended fragment size (RFS). The study defines featured traffic crash rates using three major traffic crash characteristics: number of vehicles in crash, manner of collision, and crash severity. The analysis uses K-means to cluster highway segments based on the featured crash rates (FCRs) from fragment sizes (segment lengths) ranging from 0.10 to 0.25 mile with an increment of 0.01 mile and evaluates the clustering results using their silhouette coefficients. The RSL method provides a standardized approach for selecting segment length, which benefits CFA studies. Also, crash hotspots will be identified using traffic crash count models developed based on traffic crash data aggregation for each fragment size ranging from 0.1 to 0.25 mile with an increment of 0.01 mile. See More

Project Details

Project Type: Dissertation
Project Status: Completed
End Date: December 31, 2023
UTC Funding: $15,000

Downloadable Products

Research NITC

  • Projects
  • Final Reports
  • NITC Researchers
  • Grant Funding
    • Overview
    • Requirements and Forms
    • Researcher Login
  • Curriculum: K-12 and University

 

© 2024 | National Institute for Transportation and Communities | 503-725-8545 | asktrec@pdx.edu