Collaborative and Distributed Autonomous Driving: A Game Theoretic Approach

The collaborative driving can significantly reduce the computation offloading from autonomous vehicles (AVs) to edge computing devices (ECDs) and the computation cost of each AV. However, since AVs usually have different computing capabilities and costs, the collaboration of AVs and the distribution of members in different collaborative groups directly affect the performance of collaborative driving. Therefore, how to develop an efficient collaborative autonomous driving scheme to minimize the driving cost becomes a new challenge. To this end, we propose a game theoretic approach to facilitate the collaborative and distributed autonomous driving. Specifically, we first design an architecture for AVs to make the collaborative driving decisions. With this architecture, we then design an auction game-based collaborative driving mechanism (AG-CDM) to decide the head AV and the tail AV of each group. Next, by considering the computation cost and transmission cost of each group, a coalition game-based distributed driving mechanism (CG-DDM) is developed to decide the optimal group distribution for minimizing the driving cost of each AV. Simulation results show that the proposed scheme can converge to a stable coalition structure and minimize the cost of each AV.