Driving Risk Quantification of Autonomous Vehicles Based on Drivers’ Risk Perception Experience

The quantification of driving risk is an important basis for ensuring the driving safety of autonomous vehicles. In order to make the autonomous vehicle have more accurate risk cognition ability, we propose a driving risk quantification method from the perspective of anthropomorphism, which draws on the driver’s risk perception experience. In this method, the intensity of the driver’s avoidance operation is used to reflect their perceived level of environmental danger; therefore, the GPR algorithm is used to construct driver’s avoidance operation intensity prediction model, which can predict the driver’s avoidance operation intensity based on the vehicles’ historical interaction information. Finally, an exponential model of driving risk and driver’s avoidance operation intensity is constructed to achieve real-time quantification of driving risk. The risk quantification method proposed in this paper proves advantageous for autonomous vehicles, empowering them to adopt more adaptable behavioral decisions without compromising driving safety.