Reproducing Weekly Crash Patterns Using a Minimalist Driver Error Model
summary
Microscopic SUMO simulations use an OpenStreetMap network corrections and averaged multi-year detector counts, implement a 3-second blackout driver-inattention model with mode-specific crash probabilities and run 100 simulation runs to compare hourly and weekly patterns. Results show morning and evening weekday peaks and flatter weekends closely match observed rates, but the model overpredicts turning-area collisions and underpredicts pedestrian-crossing conflicts, likely from missing pedestrian counts and unmodeled visibility/parking effects, so temporal replication is strong while local spatial detail is lacking.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 1:30Traffic safety modeling gap
- 3:58Schlesisches Tor analysis
- 8:11Driver inattention modeling
- 12:52Validation framework
- 14:40Temporal crash patterns
- 16:28Spatial validation
- 20:21Kernel density findings
- 23:15Model limitations
- 24:37Summary
related talks
14mAnalysis and Modelling of Road Traffic Using SUMO to Optimize the Arrival Time of Emergency Vehicles
16mAutomated Calibration of Traffic Demand and Traffic Lights in SUMO Using Real-World Observations
14mThe state of Bicycle Modeling in SUMO
19mSensor-based Flow Optimization on connected real-world intersections via a SUMO Feature Gap Analysis
44mSeeing Real-Time Traffic Problems
25mErlang-based Desynchronised Urban Traffic Simulation (...)
17mSignal priority for improving fluidity and decreasing fuel consumption
13mThe Effects of Route Randomization on Urban Emissions