List All Projects List Graduate Projects List Senior Design Projects List REU Projects Other Undergraduate Projects
Project # 84
Real-Time Traffic Signal State Detection Using YOLOv11 on Citywide Traffic Cameras in Arlington
| Team Lead: | Sandesh Sapkota |
| Other Team Members: | Peirong Wang |
| Faculty Advisor(s): | Taylor Li |
| Department: | CE |
| Project Type: | REU |
| Title: | Real-Time Traffic Signal State Detection Using YOLOv11 on Citywide Traffic Cameras in Arlington |
| Abstract: | The City of Arlington has established real-time access to citywide traffic monitoring cameras via a dedicated fiber connection to the ACTION Lab at the University of Texas at Arlington (UTA). This infrastructure enables advanced, data-driven traffic management initiatives. The ACTION lab at UTA developed AVAST(AI-Empowered Vehicle Analytics for Smart Transportation) to count vehicles and detect near-misses at an intersection. Although the AVAST platform currently supports real-time traffic counting, lack capabilities for traffic signal state recognition. Direct access to signal-phase data from traffic controllers remains challenging due to infrastructure restrictions and restricted access. To overcome this barrier, the ACTION Lab is developing a non-invasive, vision-based system for live detection of traffic signal states using YOLOv11, a state-of-the-art object detection model distinguished for its speed, accuracy, and efficiency in real-time applications. The system uses existing camera feeds to detect and classify signal states without requiring controller integration. A custom dataset comprising approximately 600 annotated images was collected at Arlington intersections, capturing different conditions, including varying angles and layouts. The dataset focuses on four primary classes: green, yellow, red, and permissive yellow. YOLOv11 was trained on this dataset and evaluated on live citywide camera streams. This approach delivers useful insights for evaluating signal performance, optimizing timing, reducing congestion, and enhancing urban mobility and safety in Arlington. By enabling reliable, real-time signal-state monitoring through computer vision, the system promotes broader intelligent transportation goals without invasive modifications to existing infrastructure. |
| Poster: | View Poster #84 (PDF) |
| Judges Scoring: | Submit Evaluation (judges only) |