Large Multimodal Models-based Undesignated Truck Parking Monitoring System at Rest Areas

Undesignated truck parking issues are prevalent in areas where truck parking facilities are scarce or overcrowded. When trucks park outside of dedicated spaces, they can obstruct emergency access routes, leading to public health and safety concerns, disrupt traffic flow, and increase the risk of theft. These problems are exacerbated in regions with a high demand for truck parking, such as District 8 in California, where nearly one-third of all parking incidents involve undesignated truck parking. Currently, the detection of undesignated parking relies heavily on manual enforcement, primarily through citations issued by patrol officers, which is costly and inefficient due to the significant resources required for patrols. Existing sensor-based truck parking detection systems also have less focus on undesignated parking due to lack of coverage. This project will develop an artificial intelligence (AI)-driven Large Multimodal Models (LMMs) based truck parking monitoring system that covers both designated truck parking and undesignated truck parking. It will build on existing work in the area of truck parking research, with a focus on incorporating new and innovative approaches.  Compared with traditional vision-based systems which can detect vehicles but lack the ability to interpret complex situations for undesignated truck parking, LMMs integrate both visual recognition and language interpretation to comprehend contextual information such as road signs, lane markers or surrounding environments. The research team will explore the integration of Set-of-Mark prompting with lightweight domain adaptation for LMMs, and the fine-tuned inference pipeline that takes advantage of site-specific labeled data to enable accurate, scalable truck parking monitoring for both designated and undesignated conditions. Based on data collected from the I-10 truck parking availability system through the research team’s recent project funded by Caltrans, the team will evaluate the proposed method’s effectiveness across multiple real-world parking lots under diverse visual conditions. 

Language

• English

Project

• Status: Active
• Funding: $220,000.00
• Contract Numbers:

  69A3552348329

• Sponsor Organizations:

  Center for Efficient Mobility

  1111 Rellis Parkway
  Bryan, Texas  United States  77807
• Managing Organizations:

  Office of the Assistant Secretary for Research and Technology

  University Transportation Centers Program
  Department of Transportation
  Washington, DC  United States  20590
• Project Managers:

  Ocon, Monica

• Performing Organizations:

  University of California Riverside

  Center for Environmental Research and Technology
  Riverside, CA  United States  92507
• Principal Investigators:

  Hao, Peng

• Start Date: 20260201
• Expected Completion Date: 20270731
• Actual Completion Date: 0
• USDOT Program: University Transportation Centers
• Source Data: 03-13-UCR

Subject/Index Terms

• TRT Terms: Artificial intelligence; Compliance; Monitoring; Parking; Roadside rest areas; Truck traffic
• Subject Areas: Data and Information Technology; Motor Carriers; Operations and Traffic Management; Safety and Human Factors;

Filing Info

• Accession Number: 01979476
• Record Type: Research project
• Source Agency: Center for Advancing Research in Transportation Emissions, Energy, and Health (CARTEEH)
• Contract Numbers: 69A3552348329
• Files: UTC, RIP
• Created Date: Feb 15 2026 4:44PM