New Approaches to Traffic Modeling Considering Complex Geometries and Data

Logos of project participants   Intersection from bird eye's perspective Copyright: © ISAC

Problem Definition

The traditional approach to road safety analysis - based on historical accident data - is subject to the restriction that a sufficient number of accidents must have occurred in order to identify corresponding accident blackspots and take appropriate measures. The predictive capabilities of the traditional approach are therefore limited and should be extended to include other options such as model-based predictions or replaced by traffic safety analyses with safety indicators.
Through model-based prediction, traffic modeling offers the possibility of predicting the number of accidents on the studied section of roadway. While modeling of traffic on highway sections is well advanced, models for more complex geometries (e.g., intersections), which are more risky due to changes in driving behavior, are not yet available to the same extent.

Project Objective

The aim of this project is to transfer the knowledge and data collected so far on traffic behavior on highway sections to the complex geometries of highway intersections. The newly developed models and numerical methods are to be combined with the previous results and supplemented by stochastic approaches (uncertainty) in order to finally obtain a multi-scale stochastic traffic forecast model for the entire freeway. In addition, the safety indicators are to be transferred to the intertwining traffic or developed further and completed with the fuzzy parameters.


The modeling on the three levels - microscopic, mesoscopic and macroscopic - will be performed by a complex conditioning and calibration procedure using microscopic traffic data and driving simulator studies. The microscopic traffic data on road sections and intersections will be partly collected and derived in the project or are partly available from other research projects. Stochastic parameters indispensable for model building will be derived using data from a weather station and based on the analysis of driving simulator studies and from the collected traffic data. The safety indicators will be further developed with regard to intertwining traffic, taking into account lateral movements and acceleration behavior of vehicles, and extended with stochastic parameters. In the project, a lane change model is first developed for more complex geometries, into which the safety indicators are integrated. This lane change model is then incorporated into the kinetic stochastic traffic models for road sections and intersections. The models are then validated using the microscopic traffic data and further driving simulator studies.

  Concept sketch for the determination of collision points Copyright: © ISAC