Robust two-dimensional platoon strategy for connected and automated vehicles

Platooning allows Connected and Automated Vehicles (CAVs) to operate efficiently and safely in coordinated formations. In mixed traffic with Human-Driven Vehicles (HDVs), platoon formation requires systematically generating and evaluating all feasible sequences of cooperative actions to account for HDVs’ unpredictable behaviors. Neglecting potential cooperative sequences can result in suboptimal platooning, unsafe maneuvers, or reduced energy efficiency. To address the challenge, this study firstly proposes a platoon scheme generation tree to identify all feasible platoon schemes for CAVs and guide the selection of the optimal one. Longitudinal and lateral control models are then designed to effectively manage the two-dimensional motion of CAVs toward the optimal platoon scheme. During the platooning process, the interaction between CAVs/HDVs are handled by a cooperative lane-changing strategy, incorporating gap negotiation, HDV behaviors prediction, and optimal cooperation timing for safe lateral merging in mixed traffic flow. Additionally, a two-dimensional driver model is employed to estimate HDV trajectories by a randomized kinematic intelligent driver model (KIDM). Case studies validate the proposed platoon methodology, demonstrating its effectiveness and robustness despite the influence of driving behaviors on platoon formation. The results indicate that this approach provides innovative insights and methods for implementing CAV platooning in practical mixed-traffic environments. This research indicates that a 2 % decrease in success rate of platoon formation for every 0.02 m increase in the lateral standard deviation of lateral movement of HDVs.