Virtual Platoon based CAVs Cooperative Driving at Unsignalized Intersection

The emergence of connected and automated vehicles (CAVs) poses a promising future of transportation systems. Complicated traffic conditions at intersections lead to challenges for cooperative driving of CAVs while ensuring safety and efficiency. This paper proposes a conflict-free cooperative control method for CAVs at unsignalized intersections. Based on the conflict points inside the intersection, we project each conflict vehicles cluster onto one virtual lane and construct multiple virtual platoons for ensuring safety. To maximize the intersection utilization, a packing problem is proposed for developing optimal car-following order, which is formulated as a mixed-integer linear programming (MILP). Considering the practical conditions at the intersection, we remove the redundant constraints to obtain nearly global-optimal solutions in real time. Numerical simulations validate that the proposed method saves 5.5%-9.3% total evacuation time under normal traffic flow.