Exploring the Performance of Streaming‑Data‑Driven Traffic State Estimation Method Using Complete Trajectory Data

摘要

This study aims to evaluate the performance of an extended floating car data (xFCD)-based traffic state estimation methodproposed by Seo et al. (2015), which does not rely on any strong assumptions such as Fundamental Diagram, using highresolutioncomplete trajectory data, viz. Zen Traffic Data (ZTD). Traffic state estimated by this method, considering randomlysampled trajectories of ZTD as those of probe vehicles with known penetration rates, are compared with ones obtained bycomplete ZTD by applying Edie’s generalized definitions. The variation in estimation errors and covering percentages areanalyzed for varying settings: spatiotemporal resolution and probe penetration rates.