Two-level hierarchical traffic control for heterogeneous urban networks

Field and simulation traffic data reveal that for some urban networks a well-defined Macroscopic Fundamental Diagram (MFD) exists, that provides a unimodal and low-scatter relationship between the network vehicle accumulation and outflow. Recent studies demonstrate that link density heterogeneity plays a significant role in the shape and scatter level of MFD and can cause hysteresis loops that deteriorate the network performance. This paper introduces a hierarchical perimeter flow control structure consisting of a high-level controller based on the model predictive control approach, where the prediction model is an aggregated parsimonious region-based MFD model and the plant is a detailed subregion-based MFD model. At the lower level, a feedback controller tries to maximize the outflow of critical regions by increasing their homogeneity. The hierarchical perimeter controller operates on the border between urban regions and manipulate the percentages of flows that transfer between the subregions such that the network delay is minimized and the distribution of congestion is more homogeneous. The proposed framework succeeds to increase network flows and decrease the hysteresis loop of the MFD compared to the existing perimeter controllers that are without heterogeneity controller.