Leveraging Intersection Trajectory Data for Microscopic Traffic Simulation

Abstract

High-resolution trajectory data enable more detailed simulation modelling and enhance the accuracy of traffic performance estimates. This paper integrates such data within a simulation and investigates the applicability and accuracy of a mi-croscopic traffic simulation in replicating real-world traffic conditions using data from the DLR research intersection in Braunschweig, Germany. While the results are promising, fragmented trajectories pose challenges in data conversion. To ad-dress this, a structured workflow was developed. A pre-calibration was conduct-ed with focus on driver imperfection and desired minimum headway. Compara-tive analysis underscores the need to examine and refine driving models for turn-ing, merging, and yellow light responses to improve realism. The granular data source proved valuable insights for model improvements.