Intelligent Driving Risk Assessment Based on Driving Risk Force Field Modeling

In response to the existing issues of incomplete elements in current risk assessment models, broad risk impact, and limited applicability, this paper presents a novel approach to the car-driving risk assessment model. This model comprehensively considers the speed, acceleration, heading angle, and other information of dynamic participating elements, achieving differentiated representation of front-end and back-end risks, and clearly defining the range of risk impact. Simultaneously incorporating lane markings and vehicle intelligence risk into the model, balancing comprehensiveness and applicability. Combining this model with artificial potential field path planning methods, a high-speed obstacle avoidance scenario is designed for simulation experiments. The simulation results indicate that, compared to the DSF model, the obstacle avoidance path derived from this risk assessment model effectively avoids collisions and unreachable targets. Additionally, the obstacle avoidance path length was reduced by 10.2%, and the mean curvature of the path decreased by 79.1 %, underscoring the model's feasibility and superiority in real-world applications.