Abstract
Recently, the number of traffic accidents is decreasing due to the development of advanced safety technologies. However, it is difficult to select an appropriate speed to pass the non-signalized intersection without a regulation stop sign for elderly drivers who do not drive with risk prediction. To support such drivers, this study aims to construct an expert driver model based on the risk potential field. This paper describes the method of a parameter estimation of the driver model in the situation that a driver passes through an unknown intersection. First, driving data of a person in the scenes of various intersections are acquired from the experiment in the Driving Simulator. Next, by using the driving data, the parameters of the driver model are estimated. Then, the proposed method estimates the parameters to pass the unknown intersection by analysing the relationship between the known driver parameters and the known intersection shape factors stored in the database. Finally, the accuracy of the proposed method is evaluated by comparing estimated parameters with the ground-truth parameters.
