With the development of communication and control technology, intelligent transportation systems (ITS) have received increasing attention from both industry and academia. However, plenty of studies providing different formulations for ITS depend on Master Control Center and require a high level of hardware configuration. The systematized technologies for distributed architectures are still not explored in detail. In this paper, we proposed a novel distributed cyber–physical system for connected and automated vehicles, and related methodologies are illustrated. Every vehicle in this system is modeled as a double-integrator and supposed to travel along a desired trajectory for maintaining a rigid formation geometry. The desired trajectory is generated by reference leading vehicles using information from multiple sources, while ordinary following vehicles use velocity and position information from their nearest neighbors and sensor information from on-board sensors to correct their own performance. Information graphs are used to illustrate the interaction topology between connected and automated vehicles. Edge computing technology is used to analyze and process information, such that the risk of privacy leaks can be greatly reduced. The performance scaling laws for the network with a one-dimensional information graph are generalized to networks with D -dimensional information graphs, and the results of the experiments show that the performance of the connected and automated vehicles matches very well with analytic predictions. Some design guidelines and open questions are provided for the future study.

中文翻译：

---

具有实现方法的互联和自动车辆分布式网络物理系统设计

随着通信和控制技术的发展，智能交通系统（ITS）受到了业界和学术界的越来越多的关注。但是，许多为ITS提供不同公式的研究都依赖于主控中心，并且需要高水平的硬件配置。分布式体系结构的系统化技术仍未详细探讨。在本文中，我们提出了一种用于连接和自动车辆的新型分布式网络物理系统，并说明了相关方法。该系统中的每辆车均被建模为双积分器，并假设其沿所需的轨迹行进以保持刚性地层的几何形状。期望的轨迹是由参考引导车辆使用来自多个来源的信息生成的，而普通的跟随车辆则使用来自其最近邻居的速度和位置信息以及来自车载传感器的传感器信息来校正其自身的性能。信息图用于说明已连接和自动车辆之间的交互拓扑。边缘计算技术用于分析和处理信息，从而可以大大降低隐私泄露的风险。具有一维信息图的网络的性能缩放定律被推广到具有D维信息图的网络，并且实验结果表明，连接和自动车辆的性能与分析预测非常吻合。为将来的研究提供了一些设计准则和未解决的问题。