A lane-changing decision framework for connected mixed traffic: safety condition analysis incorporating risk-coupled impact potential

Advancements in vehicle connectivity and autonomous driving technologies are shaping a connected mixed traffic environment, comprising connected autonomous vehicles (CAVs) and connected manually-driven vehicles (CMVs). To address the complex demands of lane-changing safety condition assessment in mixed traffic, this paper proposes a decision framework integrating multidimensional vehicle factors. First, a risk-coupled impact potential model is employed to quantify the comprehensive risk level posed by an adjacent vehicle to the subject vehicle. Second, through a fine-grained analysis of the safety lane-changing processes relative to adjacent vehicles at different positions, the critical-safety motion states are dissected, and the calculation methods for the inter-vehicle travel distance differences are derived. Finally, by incorporating the critical-safety distances obtained from the risk-coupled impact potential model, three minimum longitudinal safety distance lane-changing decision models are established. Numerical simulation results demonstrate that the minimum longitudinal safety distances are closely related to and reasonably influenced by vehicle speed, acceleration, driving type, and CMV driver aggression-proneness. Compared to the traditional model, the proposed framework responds more effectively to mixed traffic factors. Additionally, it demonstrates robustness and clear response characteristics to uncertainties in input parameters. This framework shows favorable potential for integration into mainstream microscopic simulation platforms and for application in real-world connected traffic scenarios, providing a safety condition decision framework for the optimized operation of intelligent transportation systems.