Takahiro Tsubota

Macroscopic Fundamental Diagram for Brisbane, Australia: Empirical Findings on Network Partitioning and Incident Detection

The macroscopic fundamental diagram (MFD) relates space mean density and flow. The existence with dynamic features was confirmed in a congested urban network in downtown Yokohama, Japan, with a real data set. Because the MFD represents areawide network traffic performance, studies have reported on perimeter control strategies and areawide traffic state estimated with the MFD concept. However, few reports of real-world examples from signalized arterial networks are available. Data are fused from multiple sources (Bluetooth devices, loop detectors, and signal phase timing). A framework is presented for the development of the MFD for Brisbane, Queensland, Australia. Existence of the MFD in the Brisbane arterial network is confirmed. MFDs (from the entire network and several subregions) are evaluated to determine the spatial partitioning to represent network performance. The findings confirm the usefulness of appropriate network partitioning for traffic monitoring and incident detection. Future research directions are addressed.

Comparative analysis of traffic state estimation ─ Cumulative counts-based and trajectory-based methods

The macroscopic fundamental diagram (MFD) relates space–mean density and flow. Because the MFD represents areawide network traffic performance, perimeter control strategies and networkwide traffic state estimation using the MFD concept have been studied. Most previous works used data from fixed sensors, such as inductive loops, to estimate the MFD, which can cause biased estimation in urban networks because of queue spillovers at intersections. To overcome this limitation, recent literature reported on the use of trajectory data obtained from probe vehicles. However, these studies were conducted with simulated data sets; few works have discussed the limitations of real data sets and their impact on variable estimation. This study compares two methods for estimating traffic state variables of signalized arterial sections: a method based on cumulative vehicle counts (CUPRITE) and one based on vehicle trajectory from taxi GPS logs. The comparisons reveal some characteristics of taxi trajectory data available in Brisbane, Queensland, Australia. The current trajectory data have limitations in quantity (i.e., the penetration rate), because of which the traffic state variables tend to be underestimated. Nevertheless, the trajectory-based method successfully captures the features of traffic states, which suggests that the trajectories from taxis can be a good estimator for networkwide traffic states.