I. Introduction
With the development of intelligent transportation system (ITS), autonomous driving technology is being employed in more and more scenarios. Advanced autonomous vehicles have the ability of perception, intelligent decision-making and path planning. It has very mature applications in industry, such as logistics distribution, public transport, etc. However, the autonomous vehicles are often self-interested in these cases, which may lead to traffic congestion [1]. Since then, cooperative driving plays an important role in promoting a harmonious traffic environment. The research on cooperative driving is mainly divided into centralized and distributed methods. The centralized methods rely on the establishment of the Internet of Vehicles (IoV). Each vehicle need to transmit information by wireless communication technology, and then the information of each node in the network is collected in the central processor for unified processing. While the distributed methods are computing independently, which are more stable but difficult to achieve the purpose of cooperation. Considering the computing limitation and the system scalability, distributed methods are more practical.