Grade Crossing Monitoring Using Deep Learning

Railway crossings are critical elements of railway safety due to the heightened risk of train collisions. The USDOT’s National Highway Traffic Safety Administration (NHTSA) reported more than 1,600 collisions between vehicles and trains in 2021, and 500 collisions at transit rail crossings in 2020. Transportation agencies and researchers are continuously working to enhance safety at railway crossings with better operating procedures and equipment to avoid accidents. Many innovative methods have been proposed to detect hazards at crossings and rail tracks using technologies such as sensors, computer vision, depth cameras, and many others. However, there is still a need to develop a holistic approach that is robust and generalizable to the many conditions and hazards related to grade crossing accidents. This project aims to investigate Artificial Intelligence (AI) and Deep Learning (DL) models to monitor grade crossings and detect various hazardous conditions such as vehicles, pedestrians, cyclists, animals, warning lights, arm positions, and others. There is a need for generalizable AI models that can be applied at different grade crossings and monitor the various conditions associated with accidents and near-miss events. To achieve that, the proposed methodology consists of (1) collecting visual data of railway crossings; (2) labeling the data for training; and (3) developing a computer vision model using deep learning that can detect hazardous conditions at railway crossings. Ultimately, the outcomes of this research support improving safety at crossings, modernizing unsafe crossings, optimizing traffic in crossings, and data sharing for research with University Transportation Center for Railway Safety (UTCRS) partners.

• Record URL:
  https://www.utrgv.edu/railwaysafety/_files/documents/research/exhibit-d_utcrs_utrgv_grade_crossing_monitoring_using_deep_learning_2023.pdf
• Supplemental Notes:
  • Partners for this project are the University of Nebraska Lincoln (UNL), the University of California-Riverside (UCR), and the University of South Carolina (UofSC).

Language

• English

Project

• Status: Active
• Funding: $65448
• Contract Numbers:

  69A3552348340

• Sponsor Organizations:

  Office of the Assistant Secretary for Research and Technology

  University Transportation Centers Program
  Department of Transportation
  Washington, DC  United States  20590

  University of Texas Rio Grande Valley

  1201 W. University Dr
  Edinburg, TX  United States  78539
• Managing Organizations:

  University of Texas Rio Grande Valley

  1201 W. University Dr
  Edinburg, TX  United States  78539
• Project Managers:

  Stearns, Amy

• Performing Organizations:

  University of Texas Rio Grande Valley

  1201 W. University Dr
  Edinburg, TX  United States  78539
• Principal Investigators:

  Ali, Gasser

  Tarawneh, Constantine

• Start Date: 20230601
• Expected Completion Date: 20240831
• Actual Completion Date: 0
• USDOT Program: University Transportation Centers Program

Subject/Index Terms

• TRT Terms: Artificial intelligence; Data collection; Machine learning; Monitoring; Railroad grade crossings; Railroad safety
• Subject Areas: Data and Information Technology; Railroads; Safety and Human Factors;

Filing Info

• Accession Number: 01898075
• Record Type: Research project
• Source Agency: University Transportation Center for Railway Safety
• Contract Numbers: 69A3552348340
• Files: UTC, RIP
• Created Date: Oct 31 2023 7:32PM