Model predictive control based on urban macroscopic fundamental diagram: A city-scale routing application

The increase in motorization creates an urgent need to improve traffic congestion management techniques in urban areas. In this context, network-level control approaches are gaining prominence, and the Macroscopic Fundamental Diagram (MFD) aggregated modeling paradigm is frequently used as a fast and reliable tool to monitor city congestion levels, particularly when the city is divided into regions, each with its own MFD curve. This paper proposes and tests a hierarchical control scheme to implement a city-level Route Guidance strategy. At the upper level, Model Predictive Control (MPC) is used to determine the optimal split ratios between regions, while at the lower level, an actuation system spreads control signals to drivers by simulating Variable Message Signs on the infrastructure. The objective of this paper is to investigate and validate the robustness of this approach. The framework is tested in a complex and realistic traffic scenario representing the entire city of Luxembourg, considering various operating conditions (e.g., parameter settings, fluctuations in travel demand), behavioral factors (e.g., drivers’ compliance) and performance indicators (e.g., system effects on both controlled and uncontrolled zones). Our results show that the control method effectively utilizes the network’s capacity, even with low driver compliance, and drives the entire network to the saturation regime, where optimal throughput is achieved. • MFD-based MPC can be effectively used for scalable route guidance applications. • MFD-based MPC can be applied in complex traffic environments (e.g. Luxembourg). • Urban traffic exhibits spatial heterogeneity and calibrated zonal MFDs capture it. • MFD-based MPC is shown to cope with various operating conditions. • MFD-based route guidance achieves benefits even with low drivers’ compliance.