A comprehensive assessment of connected and automated driving is imperative before its large-scale deployment in reality, which can be economically and effectively implemented via a credible simulation platform. Nonetheless, the key components of traffic dynamics, vehicle modeling, and traffic environment are oversimplified in existing simulators. Current traffic simulators normally simplify the function of connected and autonomous vehicles by proposing incremental improvements to the conventional traffic flow modeling methods, which cannot reflect the characteristics of the realistic connected and autonomous vehicles. On the other hand, typical autonomous vehicle simulators only focus on individual function verification in some specific traffic scenarios, omitting the network-level evaluation by integrating both large-scale traffic networks and vehicle-to-anything (V2X) communication. This paper designs a comprehensive simulation platform for conventional, connected and automated driving from a transportation cyber–physical system perspective, which tightly combines the core components of V2X communication, traffic networks, and autonomous/conventional vehicle model. Specifically, three popular open-source simulators SUMO, Omnet++, and Webots are integrated and connected via the traffic control interface, and the whole simulation platform will be deployed in a Client/Server model. As the demonstration, two typical applications, traffic flow optimization and vehicle eco-driving, are implemented in the simulation platform. The proposed platform provides an ideal and credible testbed to explore the potential social/economic impact of connected and automated driving from the individual level to the large-scale network level.

在实际大规模部署之前，必须对互联和自动驾驶进行全面评估，这可以通过可靠的仿真平台经济有效地实现。尽管如此，在现有的模拟器中，交通动态，车辆建模和交通环境的关键组成部分还是被简化了。当前的交通模拟器通常通过对常规交通流建模方法提出渐进的改进，来简化联网和自动驾驶汽车的功能，这些方法无法反映现实的联网和自动驾驶汽车的特性。另一方面，典型的自动驾驶汽车模拟器只关注某些特定交通情况下的个别功能验证，通过集成大型交通网络和车辆到任何事物（V2X）通信，省略了网络级评估。本文从交通运输网络物理系统的角度设计了一个用于常规，互联和自动驾驶的综合仿真平台，该平台紧密结合了V2X通信，交通网络和自动/传统车辆模型的核心组件。具体来说，三个流行的开源仿真器SUMO，Omnet ++和Webots通过流量控制接口集成和连接，并且整个仿真平台将部署在客户端/服务器模型中。作为演示，在仿真平台中实现了两个典型应用，即交通流优化和车辆生态驾驶。