A scalable unsignalized intersection system for automated vehicles and semi-physical implementation

With the rapid development of automated vehicle (AV) technologies, unsignalized intersection scenario has attracted a lot of research attentions in the field of intelligent transportation system (ITS). In this paper, to improve the real-time unsignalized intersection coordination performance and guarantee the reliability of vehicle scheduling, a distributed and scalable scheduling algorithm (DSA) for unsignalized intersection system is proposed. First, to make the vehicle passing the intersection practical, we present a collision-free traffic scenario which is based on the improvement of the existing unsignalized scheduling approach. Specifically, the turning directions of automated vehicles would not be restricted so that AV can turn in a practical arc curve trajectory. Second, a distributed consistency protocol (DCP) based on the edge computing of roadside units is proposed to ensure the real-time scheduling messages transfer. Finally, to verify the proposed unsignalized intersection system, a semi-physical simulation platform (SPSP) is established on a sand table model. Both the software simulation results and the sand table measured data of the SPSP have verified the effectiveness of the algorithm, which is believe to further promote the practical engineering applications of automated vehicles.