Enhancing Traffic Flow Understanding by Two-Dimensional Microscopic Models - ETF2D

• Applicants: Professor Dr. Ostap Okhrin; Dr. Martin Treiber
• Subject Area: Traffic and Transport Systems, Intelligent and Automated Traffic
• Term: since 2021
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 456691906

One of the most important limitations of current driving behavior models is the separation of car-following and lane-changing models. This includes the underlying models of commercial traffic simulation software. The resulting ignorance of finite lane-changing durations, longitudinal-lateral coupling, and cooperation is expected to lead to unrealistic traffic flow characteristics. On the other hand, the controllers of prototype autonomous vehicles treat the longitudinal and lateral dynamics in a unified way. However, since they are based on blackbox approaches (deep neuronal networks/machine learning), they are unsuitable for an understanding of real traffic flow. We propose to develop two-dimensional continuous traffic flow models that will capture the longitudinal and lateral dynamics together. The proposed model is based on a novel fully two-dimensional human driving model for non-lane based mixed traffic and ultimately on a longitudinal model such as the Intelligent-Driver Model (IDM). In a next step, we calibrate and validate the models on trajectory data that we obtain from public ressources or own fieldwork. Finally, we will incorporate the developed models in an open-source microscopic traffic simulation software in order to understand two-dimensional traffic flow and evaluate different traffic management measures. Crucial for this undertaking are the intuitive and meaningful parameters of the proposed models, particularly that of the underlying IDM. This research encompasses several fields of interest such as traffic physics, statistics, mathematics, transportation engineering, computer vision, and computer simulation.

• DFG Programme: Research Grants