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.

随着自动驾驶汽车技术的飞速发展，无信号交叉路口场景在智能交通系统（ITS）领域引起了很多研究关注。为了提高实时无信号交叉口协调性能，保证车辆调度的可靠性，提出了一种无信号交叉口系统的分布式可扩展调度算法（DSA）。首先，为了使通过交叉路口的车辆切实可行，我们提出了一种基于现有无信号调度方法的改进的无碰撞交通方案。具体地，自动车辆的转弯方向将不受限制，以使AV可以按照实际的弧形曲线轨迹转弯。第二，提出了一种基于路边单元边缘计算的分布式一致性协议（DCP），以保证调度消息的实时传输。最后，为了验证所提出的无信号交叉口系统，在沙盘模型上建立了一个半物理模拟平台（SPSP）。SPSP的软件仿真结果和沙盘实测数据均证明了该算法的有效性，相信将进一步促进自动驾驶汽车的实际工程应用。