Optimizing Signal Timing Control for Large Urban Traffic Networks Using an Adaptive Linear Quadratic Regulator Control Strategy

Wang, Hong; Zhu, Meixin; Hong, Wanshi; Wang, Chieh; Tao, Gang; Wang, Yinhai · 2020 · OpenAlex-citations

DOI: 10.1109/tits.2020.3010725

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 timing for large urban networks, aiming to reduce travel delays and energy consumption while maintaining computational feasibility. Existing methods often focus on individual intersections or rely on proprietary, black-box algorithms that are difficult to evaluate. The authors propose an adaptive multi-input, multi-output (MIMO) control strategy based on Linear Quadratic Regulator (LQR) theory. This approach explicitly models interactions between intersections and updates system parameters online to handle real-time traffic dynamics and uncertainties, offering a transparent alternative to complex heuristic or reinforcement learning methods. The study utilizes a microscopic traffic simulation environment using PTV VISSIM, calibrated with real traffic flow data from a 35-intersection network in Bellevue, Washington. The traffic system is modeled as a discrete-time dynamic system where the state variables are vehicle travel delays in North-South and East-West directions, and the control inputs are the green time durations for the North-South phases. To manage the nonlinearity of the traffic system, the authors linearize the model around operating points and employ a normalized least squares algorithm to estimate the system parameter matrices online. An adaptive LQR controller is then designed to minimize a cost function comprising both traffic delay and the magnitude of control input changes, ensuring smooth signal adjustments. The proposed adaptive LQR method was evaluated against three state-of-the-art baseline methods: max-pressure control, self-organizing traffic lights (SOTL), and independent deep Q network (IDQN) control. Simulation results demonstrated that the adaptive LQR strategy achieved shorter average traffic delays across the network compared to all three baseline methods. The study also included an ablation study to validate the necessity of the adaptive components, confirming that online parameter estimation and the specific LQR formulation contributed significantly to performance improvements over static or non-adaptive linear feedback controls. The significance of this work lies in providing a computationally efficient, transparent, and robust framework for network-wide traffic signal control. By using an adaptive LQR approach, the method effectively accounts for intersection interactions and system uncertainties without requiring large historical datasets or complex black-box logics. The results suggest that this strategy is a viable candidate for real-world deployment, offering better performance than current adaptive systems like SCOOT or OPAC, which rely on offline models or heuristic rules. The study highlights the potential of control-theoretic approaches to improve urban traffic efficiency and safety.

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.

• regulatory evaluation