Cooperative Control for Mixed Traffic at Uncontrolled Intersections Based on Leader-Follower Decision-Making Process

As mixed traffic of human-driven vehicles (HDVs) and connected autonomous vehicles (CAVs) increases, incorporating HDV unpredictability and heterogeneity into cooperative control becomes a major challenge. This paper proposes a cooperative control method for uncontrolled intersections using a leader-follower decision-making process. Vehicles' roles are assigned through a directed graph and the weighted maximum clique algorithm breaks loops while considering HDVs' influence. After loop breaking, CAVs plan trajectories using dynamic programming and quadratic planning, with results transmitted via V2X communication to a central controller. Unlike traditional graph-based methods, our model allows lower-priority vehicles to find passage windows between higher-priority vehicles, optimising interaction space. Through simulation analysis, we discussed the effects of traffic flow density and CAV penetration rate on the model's performance and compared it with the traffic light control and first-in-first-out scheduling algorithms, demonstrating that our model has a better traffic flow characteristic.