Traffic control of mixed-fleet vehicles approaching intersections with collaborative coupling

In recent years, connected and autonomous vehicles (CAVs) have become increasingly prevalent. However, they cannot fully replace human-driven vehicles (HDVs), making mixed traffic flow the dominant traffic condition. In this study, the intersection and its upstream segment are defined as the approaching area and the cooperative area, respectively. A dynamic lane-guidance algorithm (SCHME) is proposed to separate mixed traffic flows and form platoons with CAVs at the front and HDVs at the rear. Considering that CAVs can only influence the immediately following HDVs, two optimization models are developed for low and high CAV penetration rates based on vehicle composition and platoon characteristics. Furthermore, conflict and coexistence relationships are modeled using graph theory to determine the vehicle-passing sequence and evaluate model performance under different CAV penetration rates. Compared with the spanning tree model, the low-ratio model reduces the average travel time by 58%, while the high-ratio model achieves a 27% reduction, demonstrating the effectiveness and superiority of the proposed approaches.