现有自动驾驶技术面向各场景分别开发，导致其在不同场景下的决策结果易冲突、控制输入不连续，难以做到多种场景的平滑过渡，严重影响了行车安全和交通效率。具备跨场景应用功能的典型代表方法为单车自动驾驶方法和单车道队列协同控制方法，但存在单车算法成本高、在多车环境下性能受限、多车道资源利用不足等问题。车-路-云一体化融合控制系统（云控系统）连接广域车路对象形成统一整体，为大范围集中式规划决策提供了理想的平台，是解决现有自动驾驶技术跨场景应用难题的重要基础。本课题基于云控系统，提出了云支持的智能网联汽车跨场景协同决策与控制方法，依靠统一性的车云分层协同架构，解决了多车道路段和交叉路口处的多车行为冲突消解难题，实现了跨场景平滑控制。首先，设计了跨场景车云协同决策与控制架构。通过分析不同交通场景中的共性要素，建立车群运动相对坐标系，描述多车协同运动基本共性行为，将绝对坐标系中因车辆模型耦合、多车轨迹交错导致的复杂协同控制难题，分层解耦为云端相对冲突消解和车端绝对运动控制，并基于滚动时域优化的思想降低累积控制误差，保障控制精度的同时，大幅度降低了问题复杂度。其次，提出了基于车云协同的路段编队控制方法。基于车云分层协同架构，面向多车道路段场景，针对路段车群几何结构动态变化特性，提出面向多车道路段的编队控制闭环流程，开发了编队内部目标位置生成、动态指派问题建模与求解等算法，并根据车群相对运动冲突类型，设计了多车相对路径冲突消解方法，保障了车群结构切换过程中的安全性。然后，提出了基于车云协同的路口编队通行方法。基于车云分层协同架构，面向交叉路口行驶场景，提出非信控、全潮汐车道交叉路口多车协同控制方法，梳理不同车辆的行为冲突，设计车群横纵向相对位置联合指派框架，构建广义多车道编队，提出考虑目标偏好的路口编队结构调整方法，避免了车群在路段和路口处的运动冲突。最后，搭建了仿真平台和实车实验平台，开展了实验验证。结果表明，本课题提出的方法具备跨场景应用能力，在保障车辆运动安全性的同时，提升了路段和路口处编队结构切换的效率和平滑性，并降低了交通系统整体能耗。

The existing autonomous driving technologies are often developed according to certain traffic scenarios, which may result in conflicts between decision instructions and uncontinuous control commands, and can’t guarantee transition smoothness between different scenarios and influence the driving safety and traffic efficiency. Among the exsting methods, single-vehicle autonomous driving and single-lane platoon control are two typical technologies. However, there are problems like high cost, performance limitation, and insufficient utilization of multi-lane resources. Cloud control system connects multiple vehicles and roadside units to provide an ideal platform for large-scale centralized decision making and coordinated planning, and becomes an important basis for cross-scenario coordination of intelligent and connected vehicles (ICVs). In this paper, the cloud-control-supported cross-scenario coordinated decision making and control method for connected and automated vehicles is proposed, and the conflict resolution problem of multiple vehicles in multi-lane road and intersection scenarios are solved based on the general hierachicle coordination framework.Firstly, the general vehicle-cloud hierachicle coordinated decision making and control framework is designed. The general vehicular behavior in multiple traffic scenarios is analyzed, and the relative coordinate system is built to describe the basic relative motion among multiple vehicles. The complicated coordinated control problem in the geodetic coordinated system is simplified into two sub problems: the coordinated relative path planning and conflict resolution problem for the cloud, and the trajectory planning and tracking problem for the vehicles, and the rolling horizon method is utilized to reduce the accumulated long-term control deviation. This framework significantly simplifies the multi-vehicle coordinated control problem, while guaranteeing controlling accuracy.Secondly, the coordinated formation control method for multi-lane road scenarios is proposed based on the vehicle-cloud hierachicle framework. The close-loop multi-lane formation control process is proposed, according to the dynamical changing feature of multi-vehicle geometric structure. The target position in formation is generated according to the vehicular and scenario parameters, and the vehicle-target assignment problem is established and solved. The relative motion conflict among vehicles in the relative coordinate system is divided into different categories and resolve respectively, which guarantees the safety of formation structure switching process.Next, the coordinated sequencing and lane assignment method for multi-lane unsignalized intersection scenarios is proposed based on the vehicle-cloud hierachicle framework. The conflict behavior among multiple vehicles is analyzed in the multi-lane unsignalized intersections with flexible lane direction. The longitudinal sequencing and lateral lane assignment method is developed, and the general formations are established on each approaching road segment. Then, the desired geometric structure on each segment is achieved by multi-lane formation control method with target preference, which avoid motion conflicts among multiple vehicles in the intersection areas.Finally, the traffic simulation testbed and the field testbed are built to verify the performance of the proposed methods. The results indicate that the proposed method is able to be applied in multiple traffic scenarios, and improve the geometric structure switching smoothness, traffic efficiency, and overall system energy consumption, while guaranteeing collision avoidance among vehicles.