Investigation of bimodal macroscopic fundamental diagrams in large-scale urban networks: empirical study with GPS Data for Shenzhen City

This paper investigates traffic dynamics in bimodal urban networks utilizing the macroscopic fundamental diagram (MFD) and the three-dimensional macroscopic fundamental diagram (3D-MFD), which are network-level traffic flow modeling tools. Although the existence and the properties of the MFD have been extensively analyzed with field data in literature, few empirical studies examine these features of the 3D-MFDs for large-scale networks. For this work, GPS data for cars and buses running in the network of Shenzhen city in China are available for analysis and this offers a great opportunity for the investigation. Interestingly, both MFD and 3D-MFD dynamics are reflected in the data. Network partition is performed to reduce the hysteresis on the MFD and the network is split into two regions for further analysis. Then the investigation focuses on the MFD relationship for buses only. The average passenger occupancy is estimated and incorporated to generate a passenger MFD (pMFD) for buses. Moreover, bus operation on dedicated bus lanes is analyzed. Having understood traffic dynamics of cars, buses, and passengers respectively, the 3D-MFDs which illustrate the joint influence of car and bus accumulations on the global network-level traffic performance are presented. Given the scatter plot of the 3D-MFDs for the two partitioned regions, analytical approximations are provided, fitting by exponential functions. These results are promising, as they confirm the traffic features that were found from simulation-based studies in previous work.