【摘要】 车路云一体化系统(Vehicle-road-cloud Integrated System, VRCIS)作为支撑高级别自动驾驶规模化落地与智慧交通转型升级的重要技术路径，目前，相关研究多聚焦于单项关键技术突破、局部场景验证或工程示范应用，仍缺乏从人工智能赋能视角对VRCIS技术体系、演进脉络与应用实践的系统性梳理。基于此，从人工智能赋能角度出发，立足于数据驱动向生成式智能的范式变革阶段特征，系统阐述VRCIS的总体架构、关键技术与应用实践。首先，明确VRCIS的核心组件，进而阐述其智能演进机制以及关键发展阶段。其次，剖析VRCIS两个阶段技术演进范式：VRCIS-1.0(数据驱动阶段)依托多源数据融合和深度学习等方法，突破单车感知局限，完善协同机制，分层次探讨协同感知、协同决策与规划、协同控制、通信等关键技术的演进路径与突破方向；VRCIS-2.0(生成式智能阶段)采用生成式大模型等方法，实现生成式协同推演与建模，并解决复杂环境长尾问题的处理。最后，总结港口、Robotaxi、物流小巴、矿山等多场景下VRCIS的应用实践，梳理VRCIS标准体系与评价体系的建设进展。研究结果可为VRCIS的技术路线制定、标准规范建立及工程化落地提供理论支撑与参考。更多还原

【Abstract】 The Vehicle-road-cloud Integrated System(VRCIS), as a key technological pathway for supporting large-scale deployment of high-level autonomous driving and the transformation and upgrading of intelligent transportation systems. Current research predominantly focuses on breakthroughs in individual technologies, validation in localized scenarios, or engineering demonstration applications, lacking a systematic review of the VRCIS technological framework, evolutionary trajectory, and application practices from the perspective of artificial intelligence(AI) empowerment. From the standpoint of AI empowerment and grounded in the paradigm shift from data-driven to generative intelligence, this paper systematically elaborates on the overall architecture, key technologies, and application practices of VRCIS. First, it identifies the core components of VRCIS and further elucidates its intelligent evolution mechanisms and key development stages. Second, it analyzes the technical paradigm evolution of VRCIS across two stages: VRCIS-1.0(data-driven stage), which leverages methods such as multi-source data fusion and deep learning to overcome the limitations of single-vehicle perception and coordination mechanisms, exploring the evolutionary paths and breakthrough directions of key technologies including collaborative perception, collaborative decision-making and planning, collaborative control, and communication at different levels; and VRCIS-2.0(generative intelligence stage), which employs generative large models to achieve generative collaborative reasoning, modeling, and handling of long-tail problems in complex environments. Finally, it summarizes application practices of VRCIS across multiple scenarios such as ports, Robotaxi, logistics shuttles, and mines, and reviews the progress in establishing VRCIS standards and evaluation systems. This paper provides theoretical support and reference for the formulation of technical routes, establishment of standards and specifications, and engineering implementation of VRCIS.更多还原