An Improved Model of Driving Risk Field for Connected and Automated Vehicles

Ye Tian, Huaxin Pei, Jingxuan Yang, Jianming Hu, Yi Zhang, Xin Pei

Abstract: Driving risk field is regarded as an effective method to evaluate the driving safety for Connected and Automated Vehicles (CAVs). The existing driving risk field models do not fully consider the impacts of vehicle geometry and vehicle kinematics. So, the performance of the existing models needs to be further improved. In this paper, we establish a more realistic model of the driving risk field. First, vehicle geometry is incorporated into the driving risk field by building an ellipse model to appropriately reflect the vehicle shape. Second, the impacts of different moving directions of vehicles are also considered in the established model to accurately represent vehicle kinematics. Third, we design a computationally efficient function to describe the relations between potential energy and force, which lays a foundation for vehicles to make moving decisions using the driving risk field. Finally, a typical car-following scenario is designed to evaluate the performance of the proposed method with the help of Next Generation Simulation (NGSIM) dataset. Simulation results demonstrate its promising performance in describing traffic safety and driving behavior.