A study on lane-changing risk evolution of vehicle group based on potential field model in the freeway tunnel approach section

In the freeway tunnel approach section, lane-changing behaviors and transitions in the driving environment exacerbate traffic flow disruptions, increase driving risks, and lead to a higher accident rate. To this end, this study presents a method to explore the risk evolution process of lane-changing in these sections and evaluate its impact on traffic flow operations surrounding lane-changing vehicles. First, a driving risk potential field model based on the field theory, which consists of a vehicle kinetic potential field and a tunnel illumination potential field, is proposed to evaluate the driving risk. Furthermore, a “vehicle group” risk graph was constructed based on graph theory, incorporating both a node-coupling driving risk model and a topological potential entropy model. Finally, trajectory datasets were collected through naturalistic driving tests to analyze the evolution of lane-changing risk and the stability of the vehicle group. From the analysis of coupling driving risk evolution, we found that in the [100, 500) m, [500, 1000) m, and [1000, 1500) m freeway tunnel approach sections, the coupled driving risk of lane-changing vehicle (LCV) was higher than that of the other vehicles in the vehicle group. In different tunnel approach sections, LCVs that received the highest risk were from different vehicles in vehicle group. LCVs received the highest field strength from the risk potential fields of the lateral vehicle (LV), front lateral vehicle (FLV), and front vehicle (FV) in the [100, 500) m, [500, 1000) m, [1000, 1500) m tunnel approach sections, respectively. Based on the absolute value of the vehicle group topological potential entropy, we observed the resilience of the vehicle group system improved with increasing distance from the tunnel entrance. Traffic flow regained stability more quickly after lane-changing disturbances in sections farther from the tunnel entrance. This study highlights that the section closer to the freeway tunnel entrance significantly impact lane-changing risk, and it takes longer for the vehicle group to recover its stability after a lane-changing disturbance. The research results offer a theoretical and methodological foundation for enhancing traffic safety measures and developing microscopic driving behavior models for freeway tunnel approach sections.