Opportunistic Platoon Formation for Multi-Lane Highway On-Ramp Merging Considering Individual Rationality

On-ramp merging on multi-lane highways remains a key challenge in road transportation systems. By opportunistically forming platoons, connected and automated vehicles (CAVs) can navigate this area more efficiently. However, existing platoon formation algorithms often overlook individual rationality and may sacrifice individual benefits to achieve global objectives. This hinders platoon formation in social traffic environments, where participants belong to different groups or entities. Additionally, maintaining a reasonable computational burden in dense, continuous traffic flow is crucial for realistic deployment. To address these issues, this paper proposes a fast platoon formation algorithm based on an auction mechanism that incorporates individual rationality. First, we formulate the platooning formation as a predecessor selection problem with an individual rationality constraint. Second, we propose an auction-based decentralized method to solve the problem within milliseconds, while maintaining precision comparable to optimization software. Finally, this paper introduces a sampling-based motion planning method to fast generate feasible trajectories for CAVs. Extensive Monte Carlo experiments demonstrate the computational efficiency and robustness of the proposed approach across varying numbers of CAVs, traffic densities, and multiple communication failure modes. The results indicate significant improvements in traffic throughput and notable reductions in aerodynamic drag, with a solving time per step of less than 20 ms.