Traffic Signal Control for Large-Scale Urban Traffic Networks: Real-World Experiments using Vision-based Sensors

Park, Ji‐Ho; Liu, Tong; Wang, Chieh; Wang, Hong; Wang, Qichao; Jiang, Zhong‐Ping · 2024 · OpenAlex-citations

DOI: 10.1109/icca62789.2024.10591878

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This paper addresses the challenge of optimizing traffic signal control in large-scale urban networks by moving beyond simulation-based validation to real-world implementation. The authors propose a Model Predictive Control (MPC) strategy that utilizes real-time traffic flow data from vision-based sensors to minimize the number of vehicles across network links, thereby reducing travel delays and energy consumption. The research is motivated by the need to validate complex control algorithms in actual field conditions, addressing issues such as sensor limitations, communication delays, and the heterogeneity of intersection configurations. The study was conducted at 24 intersections in downtown Chattanooga, Tennessee, covering an area of approximately 0.68 square miles. The experimental setup employed GridSmart cameras with fish-eye lenses to capture real-time data, including vehicle volume, speed, and occupancy. Detection zones were manually configured and validated using volume-occupancy fundamental diagrams to ensure data accuracy. The MPC controller was designed to comply with National Electrical Manufacturers Association (NEMA) dual-ring constraints, ensuring safety and compatibility with existing infrastructure. To manage computational complexity and heterogeneity, intersections were grouped by cycle length, and the controller calculated optimal phase splits using a quadratic programming approach that minimized a cost function over a finite horizon while accounting for future states and system constraints. Experiments were performed during afternoon rush hours (3:30 pm to 5:30 pm) from March 28–30, 2023, and compared against baseline data collected the previous week (March 21–23) using standard semi-actuated NEMA controllers. Performance was evaluated using three metrics: Total Time Spent (TTS), Relative Queue Balance (RQB), and Percentage Arrival on Green (AoG). The results demonstrated significant improvements with the MPC controller. TTS decreased by 5.56%, 4.99%, and 12.6% on Tuesday, Wednesday, and Thursday, respectively. RQB decreased by 8.63%, 6.28%, and 19.5% over the same period. Additionally, AoG increased by 7.48%, 16.9%, and 6.59%, indicating better coordination and reduced stops for minor phases. The significance of this work lies in its successful deployment of an MPC-based traffic signal control system in a real-world grid network, providing a practical framework for other researchers aiming to transition from simulation to field experiments. The study confirms that MPC controllers can effectively handle complex dynamics and uncertainties, outperforming traditional methods in reducing congestion and improving traffic flow efficiency. The authors conclude that further investigation is needed to scale the approach to larger networks with more than 50 intersections and to explore decentralized control schemes for broader application.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• adaptive driving beam