Driver steering behavior model based on lane-keeping characteristics analysis

To build a driver steering model which can reflect drivers´ lane-keeping characteristics, twelve drivers´ naturalistic driving data on real inter-city highway are recorded. These data are used to analyze drivers´ general characteristics in lane-keeping process and how factors, such as lane position, disturbing objects nearby and road curvature influence drivers´ behaviors. Then, the concept of `expected lateral driving zone´ and a driver steering behavior mechanism are proposed. A driver steering behavior model is designed based on this mechanism and predictive optimal control theory. Three typical scenarios are selected to compare this new model with a single-point preview model and driver´s real steering behavior on performance. The results show that the new model can make a more accurate prediction about driver´s steering angle.