Systematic Error Correction of SUMO Traffic Simulator's HBEFA Vehicle Emission Model

This paper analyses the HBEFA-based instantaneous vehicle emissions model implemented in the SUMO microscopic traffic simulation. We identify significant discrepancies between real vehicle emissions measured with portable emission measurement systems and the modeled emissions. Corrections to the model are proposed to account for road gradients, transients, and engine management functions such as fuel cut-off, start-stop systems, and hybrid engine management. These corrections require minimal vehicle-related parameters and thresholds, making them suitable for both large-scale traffic simulation and emissions analysis of real fleet trajectory data. The proposed corrections are validated against real measurements, resulting in an accuracy gain of over 50% at the trajectory level. The network level was demonstrated on a large-scale traffic simulation. The overall results show that the baseline model is less accurate for $NOx$ and $PM$ and lacks the ability to properly account for HEVs.