Low-Cost AI-Based System for Temporary Traffic Control Review
Work-zone fatalities in the United States increased by about 50% between 2013 and 2022, with a surge of 33% in 2021 alone. To protect workers and guide drivers safely through modified traffic patterns, temporary traffic controls (TTCs) are used that must be regularly inspected for proper functioning. However, the current inspection process is manual and resource intensive, typically involving a three-person team: one person to drive, another to photograph, and a third to document observations. This staffing requirement limits the frequency and the geographic coverage of safety reviews. For NCHRP 20-30/IDEA 264, the research team will develop an open-source artificial intelligence (AI)-powered TTC inspection software that leverages Vision Language Models and requires just one inspector with a dashcam and an internet-connected computer. The inspector will drive through the work zone, upload dashcam footage for processing, and review AI-detected issues on an interactive video. The system will transform TTC reviews in several significant ways: (1) Convert a three-person operation into a streamlined, single-operator system, enabling more frequent inspections across wider geographic areas without additional labor cost. (2) Apply assessment criteria uniformly and objectively across all work zones. This offers the potential to deliver more consistent and accurate evaluations regardless of inspector fatigue, regional staffing differences, and complex work zones. (3) Harness the growing availability of crowdsourced dashcam footage from commercial fleets and autonomous vehicles. This integration transforms work-zone monitoring by ensuring remote, widespread, continuous geographical coverage, including during challenging conditions such as nighttime and adverse weather. The team will focus on high-priority deficiencies the Texas Department of Transportation identified that carry the highest penalties and represent significant safety hazards in work zones. A structured database of historical review reports will be created for use for training, validation, and prototype testing. Working with Texas Department of Transportation, an evaluation metric will be established and refined. This will be followed by a system engineering task in which core components or modules of the proposed system will be developed and tested individually and in end-to-end testing using a dataset. The system’s report generation component will synthesize component outputs into standardized inspection documentation, incorporating observations, images, regulatory citations, and location data for each identified deficiency. Finally, the AI-based system will be tested and validated across at least three active work zones in Texas that have ongoing, traditional TTC inspections.
Language
- English
Project
- Status: Active
- Funding: $149,935.00
-
Contract Numbers:
Project 20-30, IDEA 264
-
Sponsor Organizations:
National Cooperative Highway Research Program
Transportation Research Board
500 Fifth Street, NW
Washington, DC United States 20001American Association of State Highway and Transportation Officials (AASHTO)
444 North Capitol Street, NW
Washington, DC United States 20001Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC United States 20590 -
Project Managers:
Jawed, Inam
-
Performing Organizations:
Texas A&M Transportation Institute
, -
Principal Investigators:
Zhang, Zhenyu
- Start Date: 20240101
- Expected Completion Date: 0
- Actual Completion Date: 0
Subject/Index Terms
- TRT Terms: Artificial intelligence; Inspection; Monitoring; Work zone traffic control; Work zones
- Geographic Terms: Texas
- Subject Areas: Highways; Operations and Traffic Management;
Filing Info
- Accession Number: 01993241
- Record Type: Research project
- Source Agency: Transportation Research Board
- Contract Numbers: Project 20-30, IDEA 264
- Files: TRB, RIP
- Created Date: Jun 23 2026 1:44PM