【摘要】 车辆自动驾驶技术近10年进入快速发展期：单车自动驾驶技术日趋成熟；车辆群体协同控制技术正迈向工程实践；车辆自动驾驶测试验证体系逐步完善。在分析了Web of Science核心合集近15年车辆自动驾驶研究论文的基础上，对单车自动驾驶技术、多车协同与群体决策控制技术、车辆自动驾驶测试验证技术的发展历程、研究现状、尚未攻克的技术难题进行了深度分析。在技术应用方面对不同应用场景下的自动驾驶技术进行了探讨，涵盖封闭场景、半开放道路、开放道路自动驾驶技术的最新应用进展。展望未来，单车自动驾驶主要聚焦在：多模态融合的4D世界表征构建；轻量化大模型架构优化设计；融合物理常识与因果推理的自动驾驶通用认知框架；考虑动力学约束与物理机制引导的自动驾驶世界模型。多车协同与群体决策控制主要聚焦在：任务引导的集群自主调度与协同决策；数据驱动的多车动力学建模与高维状态控制；多车协同运载系统的全场景虚拟验证。车辆自动驾驶测试验证的发展趋势为：测试范式由覆盖驱动向风险驱动转变；场景工程向生成化与智能化方向演进；测试验证从工具驱动向智能驱动转变；安全验证向“可解释与可监管”方向发展。该综述可为学术界与工业界全面把握自动驾驶技术的发展历程、研究现状及未来趋势提供系统性参考与实践指引。更多还原

【Abstract】 Vehicle autonomous driving technology has entered a period of rapid development over the past decade: single-vehicle autonomous driving technology has become increasingly mature; vehicle group cooperative control technology is moving towards engineering applications; testing and verification system for vehicle autonomous driving has been gradually improved.Based on the analysis of research papers on vehicle autonomous driving retrieved from the Web of Science Core Collection over the past 15 years, this paper presents an in-depth review of the development progress, current research status, and unresolved technical challenges in single-vehicle autonomous driving, multi-vehicle cooperation and group decision-making control, as well as autonomous driving testing and verification technologies. In terms of technical applications, this study explores the latest technological progress across diverse scenarios, including enclosed scenarios, semi-open roads, and open road environments.Looking ahead, the research on single-vehicle autonomous driving will primarily focus on construction of 4D world representations via multi-modal fusion, architectural optimization of lightweight large-scale models, a general cognitive framework for autonomous driving integrating physical common sense and causal reasoning, autonomous driving world models guided by dynamic constraints and physical mechanisms. The research on multi-vehicle cooperation and group decision-making control mainly focuses on task-oriented autonomous cluster scheduling and collaborative decision-making, data-driven multi-vehicle dynamics modeling and high-dimensional state control, full-scenario virtual validation for multi-vehicle collaborative transportation systems. Furthermore, the development trends of vehicle autonomous driving testing and verification are shifting toward the testing paradigm is shifting from coverage-driven to risk-driven, scenario engineering is evolving toward generative and intelligent upgrading, testing and verification are transforming from tool-driven to intelligence-driven, safety verification is advancing toward explainability and supervisability. This review aims to provide systematic references and practical guidelines for academia and industry to comprehensively grasp the development progress, research status, and future trends of autonomous driving technology.更多还原