Modeling of drivers´ decision making of merging space on expressways

Merging into the traffic flow on an expressway is a very challenging driving task. It is important to understand driver merging behaviors for building an effective driver assistance system. The first and most important task a driver performs when merging is decision making of the merging space in the oncoming traffic on an expressway. In this study, we aim to build a mathematical model to represent the decision making of the merging space of different drivers. The proposed model is based on the logistic regression model. Further, data for the proposed model is obtained from simulation experiments. We clarified that some explanatory variables were important for the proposed model by applying stepwise method for selecting the explanatory variables. High effectiveness in predicting the merging space by the proposed model was verified by the prediction accuracy of the model. Further, individual differences in merging behaviors among drivers are considered in this model. Therefore, it is implied that a driver´s skill can be inferred by the investigation into the merging behavior of each driver by the model.