通过网联协同技术，智能网联汽车能克服单车自动驾驶的技术瓶颈，并产生新的车路云融合的汽车交通系统形态。然而，现有研究不具备利用车路云融合来实现复杂交通场景下融合感知、决策与控制的系统概念与架构。在系统闭环控制技术方面，现有研究难以实现服务于智能网联汽车的路网感知构型高效优化，难以实现路网中车辆行驶规划与交通控制的充分融合，难以同时处理不同场景下网联车辆控制对时延特征有关系统稳定性与控制性能的要求。因此，本文提出了智能网联汽车云控系统的概念与架构，及含三个环节的系统闭环控制方法，包括面向智能网联汽车的路侧感知构型设计方法、混合交通下智能网联汽车融合规划方法、时变时延下网联车辆控制方法，为车路云融合控制提供理论基础与方法支撑。 首先，提出了智能网联汽车云控系统的概念与架构。设计了基于信息物理系统理论的运行过程，提炼了五个核心特征，构建了“三层四级”的系统架构、应用架构与五个方面的关键技术架构，设计了典型应用系统闭环控制架构。 其次，提出了面向智能网联汽车的路侧感知构型设计方法。构建了传感器组感知范围模型，计算了其与弯道的几何匹配，分解了路网传感器布置优化问题，提出了实现最小总布置成本下路线感知覆盖的多类型传感器组布置优化方法，证明了方法的最优性，实现了相邻传感器组感知量测的数据关联。 接着，提出了混合交通下智能网联汽车融合规划方法。构建了考虑智能网联汽车渗透率与混合交通流特性的城市路网混合交通流宏观模型，建立了调节路网交通的智能网联汽车与交通调控方式的模型，调控方式包括路段上多车协同、路口处多车协同、多车路径协同引导、信号灯动态控制，提出了优化路网整体交通性能的智能网联汽车与交通调控统一优化方法。 然后，提出了时变时延下网联车辆控制方法。构建了考虑时延切换特征的网联车辆控制系统稳定性定理，设计了减小定理保守性的预定义参数确定方法，分别提出了面向无模型时延与马尔可夫模型时延的系统稳定性定量评价方法，建立了同时考虑时延下系统稳定性、控制误差边界与控制快速性的控制器优化方法。 最后，开展了基于仿真与实车平台的试验。结果显示所提出的方法在系统的感知构型设计、车辆融合规划与网联车辆控制方面具有可行性与先进性。

With the help of connected cooperative technologies, intelligent and connected vehicles (ICVs) can overcome the technical bottleneck of autonomous vehicle, and create a new vehicular transportation system based on the integration of vehicle, road and cloud. However, the existing studies haven't possessed the concept and architecture of a system that integrates vehicle, road and cloud and performs integrated perception, decision and control in complex traffic scenarios. And with respect to the systematic closed-loop control technology, the existing studies can hardly perform the efficient optimization of the roadside perception configuration for ICVs’ assistance, can hardly achieve the adequate integration of driving planning and traffic control in a road network, and can hardly deal with the requirements of networked vehicle control in various driving scenarios on system stability related to delay characteristics, and control performance. Therefore, this paper proposes the concept and the architecture of the ICVs’ cloud control system, and proposes the systematic closed-loop control method including the configuration design method of roadside perception for ICVs, the integrated planning method for ICVs in mixed traffic conditions, and the method of networked vehicle control under time-varying delay, which provides a theoretical foundation and methodological support for the coordination control based on vehicle-road-cloud integration. Firstly, the concept and the architecture of the ICVs’ cloud control system are proposed. The working procedure based on cyber-physical system theory is designed, the five core characteristics are summarized, the three-layer and four-level system architecture, the application architecture, and the key technological framework in five aspects are established, and the closed-loop control framework of a typical system implementation is designed. Secondly, the configuration design method of roadside perception for ICVs is proposed. The detection coverage of sensor set is modeled, and the geometric match with curve roads is calculated. The road network sensor placement problem is decomposed, and the optimization method for deploying multi-type sensor sets along a route for full detection area coverage and the minimum total deployment cost is proposed. The optimality of the proposed method is proved. And the data association of the measurements of adjacent sensor sets is implemented. Thirdly, the integrated planning method for ICVs in mixed traffic conditions is proposed. The road network mixed traffic flow macroscopic model considering ICV penetration rate and the dynamics of mixed traffic flow dynamics is established, and the measures related to ICVs and traffic to regulate road network traffic flow are modeled which include multiple vehicle cooperation on a road segment, multiple vehicle cooperation at an intersection, cooperative route guidance and traffic light control. And the optimization method for integrated ICV and traffic regulation that optimizes the overall performance of road network traffic is proposed. Fourthly, the method of networked vehicle control under time-varying delay is proposed. The system stability theorem considering the characteristics of delay switches is established for networked vehicle control system. The procedure for determining predefined parameters is designed to reduce the conservativeness of the theorem. The system stability quantitative evaluation methods for unmodeled delay or Markov model delay are proposed respectively. And the controller optimization method considering system stability under delay, control error boundary and control rapidity is established. Finally, experimental validation is conducted based on simulation platform and vehicle testbed. Results show the feasibilities and advantages of the proposed methods of the system in perception configuration design, vehicle integrated planning and networked vehicle control.