大语言模型技术在交通运输领域的应用研究

Research on the Application of Large Language Model Technology in the Transportation Sector

摘要

本研究综述大语言模型在交通运输领域的应用现状，探讨大语言模型在该领域的可行性和价值，给出面向交通运输行业的 专用大语言模型构建思路。通过分析近年有关交通运输大语言模型宏观政策、产业案例和学术研究现状，构建基于检索增 强架构的交通运输大语言模型，阐明基础功能层、技术核心层、智能交互应用层等各层面的内涵，给出融合行业知识的数 据准备、预训练、知识训练、微调、测试和部署一站式流程。研究发现，大语言模型在交通运输领域具备较高的实用价 值，检索增强架构和行业知识微调方法可以实现交通运输大语言模型的实时查询、自动报表及决策支持等，降低数据跨系 统使用门槛，使传统需要数个小时的资料抽取、整理变成秒级响应，也能自动生成画像和处置意见，减轻人工负担，可有 效提升交通运输管理效率。

Abstract

This study reviews the current applications of large language models in the transportation sector, examines their feasibility and value in this field, and proposes a specialized framework for developing large language models tailored to transportation needs. By analyzing recent macro-level policies, industry case studies, and academic research on transportation large language models, the study constructs a retrieval-enhanced architecture for such models. It clarifies the functional components—including the foundational layer, technical core layer, and intelligent interaction application layer—and provides a streamlined workflow encompassing data preparation, pre-training, knowledge refinement, fine-tuning, testing, and deployment integrated with industry-specific knowledge. The findings demonstrate that large language models hold significant practical value in transportation. The retrieval-enhanced architecture and industry knowledge fine-tuning methods enable real-time queries, automated reporting, and decision support, reduce barriers to cross-system data utilization, transform time-consuming data extraction and processing into seconds-long operations, and automatically generate user profiles and action recommendations—significantly alleviating manual workload and enhancing transportation management efficiency.

关键词

交通运输 / 大语言模型 / 检索增强生成 / 数据查询 / 模型微调

Key words

transportation / large language models / retrieval-enhanced generation / data query / model fine-tuning

参考文献

[1] 马童, 杜永林, 孙佳露, 等. 大语言模型驱动的铁路信息风险隐患智能识别设计与实现[J/OL]. 集成技术,1-14[2026-05-20]. https://link.cnki.net/urlid/44.1691.T.20260319.1745.002.

[2] 孟庆伟, 钟宁桦, 郝雨桐, 等. 非结构化数据增强出口预测: 基于机器学习与大语言模型[J]. 世界经济, 2026, 49(03):3-35. DOI:10.19985/j.cnki.cassjwe.2026.03.002.

[3] 张立安, 钱英军, 陈公兴. 大语言模型在物流领域发展的应用研究[J]. 物流科技, 2026, 49(03):45-47+52.DOI:10.13714/j.cnki.1002-3100.2026.03.011.

[4] 张丽岩, 范敏祎, 马健. 基于大语言模型的交通拥堵预测研究[J]. 物流科技, 2025, 48(17):90-93.DOI:10.13714/j.cnki.1002-3100.2025.17.024.

[5] 张丽, 郑晓峰, 刘礼勇, 等. 大模型在交通运输行业的应用路径[J]. 交通运输研究, 2025, 11(04):67-78.DOI:10.16503/j.cnki.2095-9931.2025.04.006.

[6] 马健, 崔玉玉, 张丽岩. 大语言模型融合知识图谱的交通事故问答系统研究[J]. 电脑与信息技术, 2025, 33(04):108-111. DOI:10.19414/j.cnki.1005-1228.2025.04.010.

[7] 郭玥君. 基于深度学习和大模型的交通流量预测与交通问答系统[D]. 大连交通大学, 2025.DOI:10.26990/d.cnki.gsltc.2025.000557.

[8] 王志峰. 面向交通领域的法律大语言模型训练微调部署技术[D]. 上海第二工业大学, 2025.DOI:10.27916/d.cnki.ghdeg.2025.000136.