Stochastic Modeling of Decision Making in Human Driving Behavior

Modeling of driver’s decision making characteristics is one of the most significant issues for design of harmonious driving assist systems. This paper presents a new strategy to build the mathematical model of the driver’s decision making characteristics based on the logistic regression model, which represents the relationship between a binary output and some continuous input variables together with an occurrence probability. Then, this model is applied to the driving data in the situation of ‘turning right’, and its usefulness is verified. The proposed model is highly expected to be used for the prediction of the driver’s decision making.