A Dynamic Traffic Calibration Framework for SUMO Based on Historical Real-World Speed Data
summary
A framework converts TomTom Vector Flow Tiles (PBF) decoded into PostGIS into time-dependent lane speeds in SUMO, assigning speed-based calibrators that act as per-lane speed signs updated every 10 minutes in 5 km/h steps. Network comes from OSM, population uses random trips and route sampler, and metrics compare simulated versus observed velocities with mean absolute error and RMSE. Proof-of-concept on Esslingen shows significant realism improvement and an acceptable mean absolute error compared to uncalibrated runs.
AI-written summary · about