Identifying and Analyzing Pass-by Crashes for the Purpose of Designing Proper Intervention Measures to Mitigate Crashes Involving Rural Population

The United States Census Bureau reports that rural areas cover about 97% of the nation’s land area and are a home to about 60 million people. About 19% of the American population lives in the rural area according to the Census Bureau. Although only 19% of the population lives in rural areas more than 70% of the 4 million miles of roadways in the United States are in rural areas. According to the NHTSA (2021) the fatality rate was 1.5 times higher in rural areas than in urban areas of the US. In Florida, the fatality rate per 100 million VMT (Vehicles Miles Travelled) in rural areas and urban areas were 2.06 and 1.64, respectively, giving a rural to urban fatality rate ratio of about 1.3. Similar to many other states of the country, on average, the road travel in rural areas in Florida is dangerous and riskier with respect to fatal accidents. This research focuses on analyzing pass-by crashes in rural areas, particularly in the Florida Department of Transportation (FDOT) District 3 (Northwest Florida). The primary goal is to identify trends and factors contributing to these crashes and propose interventions aimed at improving transportation safety for rural populations. By focusing on rural transportation, the study aligns with the broader objective of promoting equity and safety in regions that are often underserved in terms of infrastructure and access to transportation resources. The project will explore innovative machine learning and statistical methods to analyze the complex interactions between drivers' social characteristics and roadway features that influence the frequency and severity of rural pass-by crashes. The findings will inform the development of countermeasures to mitigate risks posed by transportation systems, particularly for vulnerable populations. Data needed to train the models were sourced from the Florida Traffic Safety Dashboard (Signal 4 Analytics), FDOT Geographic Information System (GIS) Open Data Hub and US Census, focusing on crash events, roadway characteristics and driver demographics. To classify pass-by crashes, distances between crash locations and the drivers' home ZIP codes were calculated, with a threshold of 30 miles used to define a pass-by crash. Logistic regression and Random Forest models were used to analyze the factors influencing these crashes, with variables such as functional class, weather conditions, vision obstruction, and type of shoulder playing significant roles in predicting crash likelihood. Preliminary results indicate that certain factors, like severe crosswinds, paved shoulders, and specific road classifications, increase the probability of pass-by crashes. However, the dataset is imbalanced, with far fewer pass-by crashes than non-pass-by crashes, and methods like under-sampling or over-sampling are being considered to address this issue. Additional future work will focus on refining the models, incorporating additional demographic and roadway data, and further validating the findings.

Language

• English

Project

• Status: Active
• Funding: $73,866.00
• Contract Numbers:

  69A3552348321

• Sponsor Organizations:

  Office of the Assistant Secretary for Research and Technology

  University Transportation Centers Program
  Department of Transportation
  Washington, DC  United States  20590
• Managing Organizations:

  Florida A&M University, Tallahassee

  404 Foote/Hilyer
  Tallahassee, FL  United States  32307
• Project Managers:

  Moses, Ren

• Performing Organizations:

  Florida A&M University, Tallahassee

  404 Foote/Hilyer
  Tallahassee, FL  United States  32307

  Cleveland State University

  Euclid Avenue at 24th Street
  Cleveland, Oh  United States  44115

  University of Washington Tacoma

  1900 Commerce Street
  Tacoma, Washington  United States 
• Principal Investigators:

  Vanli, Omer Arda

  Moses, Ren

  Kidando, Emmanuel

  Kitali, Angela

• Start Date: 20240601
• Expected Completion Date: 20250531
• Actual Completion Date: 0
• USDOT Program: University Transportation Centers Program

Subject/Index Terms

• TRT Terms: Crash causes; Crashes; Machine learning; Passing; Rural areas; Statistical analysis; Traffic safety
• Geographic Terms: Florida
• Subject Areas: Data and Information Technology; Highways; Safety and Human Factors;

Filing Info

• Accession Number: 01945639
• Record Type: Research project
• Source Agency: Rural Equitable and Accessible Transportation Center
• Contract Numbers: 69A3552348321
• Files: UTC, RIP
• Created Date: Feb 12 2025 5:48PM