| 37 | 0 | 126 |
| 下载次数 | 被引频次 | 阅读次数 |
当前,以大模型为代表的生成式人工智能正推动交通认知与治理范式的深刻变革。中国高速公路网作为全球规模最大的超大规模复杂交通系统,传统依赖局部感知和分散决策的管理模式已难以适应发展需求;而各地交通大模型建设在缺乏统一顶层设计的情况下呈碎片化发展态势,存在形成新的“智能孤岛”风险。本文提出,中国交通智能化建设亟需实现从“局部智能”向“系统智能”的范式跨越。在此基础上,系统阐述了构建全国统一协同的 AI 交通大模型体系的战略紧迫性、顶层架构与落地路径。在体系构想层面,提出以“感知⁃认知⁃决策⁃执行”一体化闭环为核心范式,以“国家统筹、省域落地”两级协同为组织架构,以“大模型做认知、小模型做执行”分层共生为能力生态的顶层设计方案;在落地实施层面,提出标准先行筑牢制度基础、场景驱动牵引价值闭环、生态共建汇聚创新合力的关键路径。研究旨在为中国交通治理能力现代化提供理论引领,为交通强国战略实施提供决策参考。
Abstract:At present, generative artificial intelligence (AI) represented by foundation models is driving a profound transformation in transportation cognition and governance paradigms. As the world's largest large-scale and complex transportation system, China's expressway network can no longer be effectively managed through traditional approaches that rely on localized perception and decentralized decision-making. Meanwhile, the development of transportation foundation models across different regions has shown a fragmented pattern in the absence of a unified top-level framework, posing the risk of creating new “intelligent islands.” This paper argues that China's transportation intelligence development urgently requires a paradigm shift from local intelligence to system intelligence. Based on this perspective, the study systematically elaborates on the strategic urgency, top-level architecture, and implementation pathways for building a nationally unified and coordinated AI foundation model system for transportation. In terms of system design, an integrated closed-loop paradigm of perception –cognition-decision-making-execution is proposed as the core framework, supported by a two-level organizational structure of national coordination and provincial implementation, and a layered capability ecosystem characterized by foundation models for cognition and lightweight models for execution. In terms of practical implementation, the paper further proposes three key pathways, that is, establishing institutional foundations through standards-first governance, driving value realization through scenariooriented applications, and fostering innovation through collaborative ecosystem development. This study aims to provide theoretical guidance for the modernization of China's transportation governance capacity and offer policy references for the implementation of the national strategy of building a transportation powerhouse.
[ 1] DWIVEDI Y K, KSHETRI N, HUGHES L, et al.Opinion Paper:“So what if ChatGPT wrote it?” Multidisciplinary perspectives on opportunities, challenges and implications of generative conversational AI for research, practice and policy[J]. International Journal of Information Management,2023,71:102642.
[ 2] BOMMASANI R, HUDSON D A, ADELI E, et al.On the opportunities and risks of foundation models[EB/OL].(2021-08-16)[2026-06-20]. https://arXiv.org/abs/2108.07258.
[ 3] HASSAN M, KABIR M E, JUSOH M, et al. Large language models in transportation: a comprehensive bibliometric analysis of emerging trends, challenges, and future research[J]. IEEE Access,2025,13:132547-
132598.
[ 4] ZHANG Z J, SUN Y J, WANG Z P, et al. Large language models for mobility analysis in transportation systems: a survey on forecasting tasks[J]. Transportation Research Record: Journal of the Transportation Re⁃
search Board,2026,2680(2):756-774.
[ 5] 中 共 中 央 国 务 院 . 交 通 强 国 建 设 纲 要[EB/OL].(2019-09-19)[2026-06-01]. https://www. gov. cn/zhengce/2019-09/19/content_5431432.htm.
[ 6] 交通运输部,国家发展改革委,工业和信息化部,等 .关 于“ 人 工 智 能 + 交 通 运 输 ”的 实 施 意 见[EB/OL].(2025-09-26)[2026-06-01]. https://xxgk.mot.gov.cn/2020/jigou/kjs/202509/t20250925_4177256.html.
[ 7] 交通运输部 . 2024 年交通运输行业发展统计公报[EB/OL]. (2025-06-12)[2026-06-01]. https://xxgk. mot.gov.cn/2020/jigou/zhghs/202506/t20250610_4170228.html.
[ 8] 陆化普,李瑞敏 . 城市智能交通系统的发展现状与趋势[J]. 工程研究-跨学科视野中的工程,2014,6(1):6-19.
[ 9] 陆化普,孙智源,屈闻聪 . 大数据及其在城市智能交通系 统 中 的 应 用 综 述[J]. 交 通 运 输 系 统 工 程 与 信 息 ,2015,15(5):45-52.
[10] LIU J L, YU W P, YANG M M, et al. Generative AI for transportation safety and resilience: a comprehensive review from a lifecycle perspective[J]. Safety Science,2026,196:107090.
[11] 中 国 移 动 . AI+ 智 慧 城 市 安 全 解 决 方 案 白 皮 书[J].(2024)[2026-06-01]. 中国信息安全,2024(10):92.
[12] NIE T, SUN J, MA W. Exploring the roles of large language models in reshaping transportation systems: a survey, framework, and roadmap[J]. Artificial Intelligence for Transportation,2025,1:100003.
[13] 肖建力,邱雪,张扬,等 . 交通大模型综述[J]. 交通运输工程学报,2025,25(1):8-28.
[14] 陆 化 普 . 智 能 交 通 系 统 主 要 技 术 的 发 展[J]. 科 技 导报,2019,37(6):27-35.
[15] 杨晓光,胡仕星月,张梦雅 . 智能高速公路交通应用技术 发 展 综 述[J]. 中 国 公 路 学 报 ,2023,36(10):142-164.
[16] 中 共 中 央 国 务 院 . 国 家 综 合 立 体 交 通 网 规 划 纲 要[EB/OL].(2021-02-24)[2026-06-01]. https://www.gov.cn/zhengce/2021-02/24/content_5588654.htm.
[17] LIU C X, YANG S, XU Q X, et al. Spatial-temporal large language model for traffic prediction[C]//Proceedings of 25th IEEE International Conference on Mobile Data Management (MDM). IEEE,2024:31-40.
[18] MAO Y C, ZHOU H L, CHEN L, et al. A survey on spatio-temporal prediction: from transformers to foundation models[J]. ACM Computing Surveys, 2026,58(4):1-36.
[19] JIN K, WI J, LEE E, et al. TrafficBERT: pre-trained model with large-scale data for long-range traffic flow forecasting [J]. Expert Systems with Applications,2021,186:115738.
[20] 李克强,戴一凡,李升波,等 . 智能网联汽车(ICV)技术 的 发 展 现 状 及 趋 势[J]. 汽 车 安 全 与 节 能 学 报 ,2017,8(1):1-14.
[21] YU B, YIN H T, ZHU Z X. Spatio-temporal graph convolutional networks: a deep learning framework for traffic forecastin[C]//Proceedings of the TwentySeventh International Joint Conference on Artificial Intelligence. International Joint Conferences on Artificial Intelligence Organization,2018:3634-3640.
[22] WU Z H, PAN S R, LONG G D, et al. Graph WaveNet for deep spatial-temporal graph modeling
[C]//Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence. International Joint Conferences on Artificial Intelligence Organization,2019:1907-1913.
[23] LIM B, ARıK S Ö, LOEFF N, et al. Temporal Fusion Transformers for interpretable multi-horizon time series forecasting[J]. International Journal of Forecasting,2021,37(4):1748-1764.
[24] ZHOU H Y, ZHANG S H, PENG J Q, et al. Informer: beyond efficient transformer for long sequence time-series forecasting[J]. Proceedings of the AAAI Conference on Artificial Intelligence,2021,35(12):11106-11115.
[25] 王江锋,舒玉东,李云飞,等 . 智慧公路数字化转型中AI 大 模 型 的 创 新 应 用[J]. 交 通 运 输 研 究 ,2025,11(4):93-103.
基本信息:
引用信息:
[1]翁孟勇,冉斌,谭现锋,等.从局部智能到系统智能:中国高速公路 AI 交通大模型体系构建的战略与治理架构[J],2026(04):1-8.
2026-07-15