Macroscopic Traffic Flow Model Calibration for Lane-free Automated Vehicle Traffic

A novel traffic paradigm was recently proposed for automated vehicles. It is characterized by two integrated principles: lane-free traffic and vehicle nudging. The latter suggests that vehicles can be influenced (nudged) by other vehicles in their vicinity, even behind them. Various vehicle movement strategies have been designed for this concept that can be employed for microscopic traffic simulation. On the other hand, macroscopic models may be used to produce large-scale simulation results with low computational effort, but also to gain insights on the emerging macroscopic properties of lane-free traffic with vehicle nudging. Furthermore, macroscopic models can be used for various significant traffic engineering tasks, including estimation and control strategy design. This paper employs an ad-hoc vehicle movement strategy for Connected and Automated Vehicles (CAVs) driving in a lane-free highway environment, to produce data that can be utilized to calibrate macroscopic models. An extended version of a well-known first-order traffic flow model is calibrated to fit the data, and optimal model parameters are determined for different highway widths and different vehicle nudging levels. The traffic conditions observed in the microscopic lane-free CAV traffic are reproduced with sufficient accuracy. Highway capacity and critical density are found to increase linearly with the increase of the highway width; and further increase of the same quantities is observed for increased levels of vehicle nudging. On the other hand, free speed is independent of the highway width and increases slightly with vehicle nudging.