A Coordinated Adaptive Signal Control Method Based on Queue Evolution and Delay Modeling Approach

Abstract

Our research offers practical solutions with direct real-world applicability: (1) Intersection-Level Optimization: We propose a novel signal control method that addresses three critical yet overlooked challenges in existing studies: (1) impact of pedestrian stages and overlap phases on signal optimization models, (2) coupling effects of signal cycles and queue lengths, and (3) stochastic vehicle arrivals in undersaturated conditions. (2) Deployable System Architecture: A cloud–edge–terminal framework has been implemented in real-world settings (with equipment brands detailed in the paper). The cloud platform provides traffic managers with an interactive interface for system monitoring and control. (3) Validation Platform: Our hardware-in-the-loop simulation system has supported multiple editions of the Shanghai Intelligent New Energy Vehicle Big Data Competition. (4) Field Results: Real-world tests on Chengaodadao, Conghua District, Guangzhou, China demonstrate a 50% reduction in stops and 27% shorter travel times in coordinated directions.

Coordinated adaptive signal control is a proven strategy for improving traffic efficiency and minimizing vehicular delays. First, we develop a Queue Evolution and Delay Model (QEDM) that establishes the relationship between detector-measured queue lengths and model parameters. QEDM accurately characterizes residual queue dynamics (accumulation and dissipation), significantly enhancing delay estimation accuracy under oversaturated conditions. Secondly, we propose a novel intersection-level signal optimization method that addresses key practical challenges: (1) pedestrian stages, overlap phases; (2) coupling effects between signal cycle and queue length; and (3) stochastic vehicle arrivals in undersaturated conditions. Unlike conventional approaches, this method proactively shortens signal cycles to reduce queues while avoiding suboptimal solutions that artificially “dilute” delays by extending cycles. Thirdly, we introduce an adaptive coordination control framework that maintains arterial-level green-band progression while maximizing intersection-level adaptive optimization flexibility. To bridge theory and practice, we design a cloud–edge–terminal collaborative deployment architecture for scalable signal control implementation and validate the framework through a hardware-in-the-loop simulation platform. Case studies in real-world scenarios demonstrate that the proposed method outperforms existing benchmarks in delay estimation accuracy, average vehicle delay, and travel time in coordinated directions. Additionally, we analyze the influence of coordination constraint update intervals on system performance, providing actionable insights for adaptive control systems.