Queue-Based Traffic Signal Control Considering Vehicle Delay Balancing for Urban Road Networks

Abstract

Traffic signal control is effective in mitigating network congestion. However, most network signal control strategies ignore the fairness of phase service at individual intersections, resulting in excessive delays for low-volume movements, particularly under unbalanced demand patterns. Although some multi-objective network control methods have been proposed to address the fairness-related issues, reweighting the objectives under different traffic conditions requires significant computational effort. This study proposes a novel queue-based Model Predictive Control (MPC) framework that incorporates delay balancing into signal control optimization for road networks. Built upon an extended segment-level flow propagation model, the proposed MPC predicts queue length and flow dynamics without relying on a sophisticated spatiotemporal discretization of road segments. A unified MPC formulation is then developed, which incorporates a virtual queue mechanism that simultaneously minimizes turn-level queue lengths and enhances delay equity across different vehicle movements. Simulation results on both a single intersection and a network demonstrate the superiority of the proposed MPC compared to existing rival methods, and it is robust against flow prediction errors.