Modeling the Effects of Motorway Traffic Control on Driving Behavior in a Microscopic Traffic Simulation
Line control systems on motorways contribute to improving traffic safety and to mitigating traffic breakdowns by means of variable speed limits, lane signals, passing restrictions and warnings. Such control measures are mostly triggered automatically in response to the prevailing traffic and weather situation. Contrary to many other types of traffic control, line control systems have only very rarely been analyzed in a microscopic traffic simulation. One major issue is the calibration of the effects certain variable message sign states have in a given situation. On the one hand, multiple messages and control strategies may overlap, making it difficult to distinguish their individual effects. On the other hand, surrounding traffic and weather conditions must be considered, as well. This paper presents a new approach to model the effects of line control systems on the driving behavior of individual vehicle-driver units in a microscopic traffic simulation. Various influencing factors as well as driver model parameters are modeled as state variables (nodes) of a Bayesian Network, which is trained based on field data from multiple motorway sections across Germany that are equipped with line control systems. This paper describes the modeling methodology, including the calibration and validation process. Furthermore, this paper describes how the model interacts with the microscopic traffic simulation at runtime, and discusses potential use cases of this approach.