首届国际工程智能大赛
今日启动
赛题解读抢先看
交通工程赛道：
面向高韧性交通系统的智能粒流协同挑战赛
Traffic Engineering Track：Intelligent Micro-Macro Coordination Challenge for High-Resilience Transportation Systems
在“交通强国”与全球智慧城市建设的浪潮下，现代交通系统的数字化与智能化转型正在加速。构建一个能够实时感知、精准预测、高效协同的智能交通体系，是解决城市拥堵、提升运行韧性的基础。此体系中的核心，便是实现从“车路分割”到“车路云一体化”的跨越。
Amid the wave of "Transportation Power" strategy and global smart city construction, the digital and intelligent transformation of modern transportation systems is accelerating. Building an intelligent transportation system capable of real-time perception, accurate prediction, and efficient coordination is the foundation for solving urban congestion and enhancing operational resilience. The core of this system lies in realizing the leap from "separated vehicle-road operation" to "vehicle-road-cloud integration".
然而，在当前的智慧交通实践中，车辆的智能驾驶与道路的交通管控长期被视为两个独立的领域，其协同控制环节是制约系统整体效能提升的主要瓶颈。传统的交通信号、可变限速等控制手段，本质上是对宏观交通“流”的被动式、粗粒度调控，占据了系统优化的主导地位，却未能充分利用智能网联汽车（ICV）作为“智能颗粒”的主动性与潜力。高质量的协同是保证系统最优的前提，而协同的缺失可能导致控制指令冲突、系统资源浪费，甚至在突发事件下引发连锁拥堵，给城市交通带来巨大的安全与效率隐患。
However, in current smart transportation practices, intelligent vehicle driving and road traffic management have long been regarded as two independent fields, and their coordinated control is a major bottleneck restricting the improvement of overall system efficiency. Traditional control methods such as traffic signals and variable speed limits are essentially passive and coarse-grained regulation of macro traffic "flow", which remains the predominant approach for system optimization but fails to fully utilize the initiative and potential of Intelligent Connected Vehicles (ICVs) as "intelligent particles". High-quality coordination is a prerequisite for optimal system performance; the lack of coordination may lead to conflicting control instructions, waste of system resources, and even cascading congestion in emergencies, posing significant safety and efficiency risks to urban traffic.
这一挑战在应对大规模、高动态的城市交通场景中尤为突出。以典型的城市快速路与匝道汇流区为例，对其进行精细化的协同管控，是保障路网稳定运行的核心步骤。管控不仅需要应对日常通勤高峰的潮汐车流，更需具备应对交通事故、恶劣天气等极端扰动工况下的快速恢复能力。这些高韧性的协同策略，是未来超大城市交通管理与应急响应决策的关键依据。然而，面对海量智能体交互、交通状态瞬息万变的任务，现有的分离式控制方法难以实现全局化和前瞻性的干预，其效果高度依赖于分割的子系统，整体效率和鲁棒性难以保证。
This challenge is particularly prominent in addressing large-scale, highly dynamic urban traffic scenarios. Taking the confluence area of a typical urban expressway and ramp as an example, refined coordinated management is a core step to ensure the stable operation of the road network. The management needs to not only cope with tidal traffic flows during daily commuting peaks but also possess rapid recovery capabilities under extreme disturbance conditions such as traffic accidents and severe weather. These high-resilience coordination strategies are key bases for future megacity traffic management and emergency response decisions. However, facing tasks involving massive agent interactions and rapidly changing traffic conditions, existing separate control methods struggle to achieve global and forward-looking interventions.Their effectiveness is highly dependent on segmented subsystems, making it difficult to guarantee overall efficiency and robustness.
人工智能技术为此提供了全新的解决思路。我们能否开发一种算法，使其不仅能理解宏观交通流的演化规律，更能洞察关键车辆（“粒”）的行为对整个系统（“流”）的蝴蝶效应，从而自适应地生成最优的车路协同策略？即通过对少数关键车辆的精准路径引导和驾驶行为“微操”，结合对关键路侧设施的“智控”，最终在保证安全的前提下，以最小的干预成本，实现系统整体运行效能的最大化。
Artificial intelligence technology provides a brand-new solution. Can we develop an algorithm that not only understands the evolutionary patterns of macro traffic flow but also gains insight into the butterfly effect of key vehicles (“particles”) on the entire system (“flow”), thereby adaptively generating optimal vehicle-road coordination strategies? That is, through precise path guidance and driving behavior "micro-control" of a few key vehicles, combined with "intelligent control" of key road-side facilities, ultimately maximizing the overall operational efficiency of the system with minimal intervention costs under the premise of ensuring safety.
这正是本项挑战赛的核心任务。参赛者需要开发一个AI协同控制算法，该算法能直接处理一个给定的复杂交通场景，并自主生成一套包含少数“智能体”车辆动态驾驶策略和关键路侧设施（如信号灯、可变限速牌）调控指令的协同方案。我们将通过一套标准化的后续交通仿真流程，从系统效率（如平均通行时间、吞吐量）、系统韧性（如拥堵恢复速度）和干预成本（如受控车辆数、能耗）等多个维度，对算法生成的协同方案进行综合性能评估。
This is precisely the core task of this challenge. Participants need to develop an AI coordinated control algorithm that can directly process a given complex traffic scenario and independently generate a coordination plan including dynamic driving strategies for a few "agent" vehicles and control instructions for key road-side facilities (such as traffic lights, variable speed limit signs). Through a standardized subsequent traffic simulation process, we will comprehensively evaluate the performance of the coordination plan generated by the algorithm from multiple dimensions, including system efficiency (e.g., average travel time, throughput), system resilience (e.g., congestion recovery speed), and intervention cost (e.g., number of controlled vehicles, energy consumption).
本挑战赛的目标是：研发能够服务于真实城市交通场景的智能化协同控制方案，将交通管理从“被动响应”升级为“主动引导”，推动智慧交通走向高效、安全与深度协同的新阶段。
The goal of this challenge is to develop an intelligent coordinated control plan that can serve real urban traffic scenarios, upgrade traffic management from "passive response" to "active guidance", and promote smart transportation to a new stage of efficiency, safety, and in-depth coordination.
