军事知识服务平台
步行与非机动车交通知识图谱系统构建与应用
Construction and Application of a Knowledge Graph System for Pedestrian and Non-Motorized Transportation
摘要 针对当前步行与非机动车交通领域多源数据关联表达不足和图数据库使用门槛较高的问题,开展了城市步行与非机动车交通知识图谱系统构建与应用研究.以Neo4j图数据库为基础,整合GIS矢量路网、步行与非机动车设施调查、街景照片及设计方法文档等多源数据,构建了包含道路步行与非机动车数据子图与设计方法子图的知识图谱,并通过知识推理完善图谱结构,建立路段与设计方法之间的关联.系统通过术语映射、大语言模型参数解析及本地查询模板,实现了自然语言查询功能,并设计可视化界面直观展示知识查询结果.应用分析表明,系统可完成路段画像分析、短板筛查、问题溯源、设计方法推荐及"评价—诊断—治理"全流程自动化查询.研究表明,该知识图谱系统能够将道路、指标、问题与设计方法组织为具有可解释性的逻辑结构,为人行道与非机动车道的精细化治理提供有效支撑.
Abstract:
To address the current issues of insufficient representation of relationships among multi-source data and the high barrier to entry for graph databases in the field of pedestrian and non-motorized transpor-tation,this paper investigates the construction and application of a knowledge graph system for urban pe-destrian and non-motorized transportation.Based on the Neo4j graph database,the paper integrates multi-source data—including GIS vector road networks,surveys of pedestrian and non-motorized transportation facilities,street-view photographs,and design methodology documents—to construct a knowledge graph comprising subgraphs for road-based pedestrian and non-motorized transportation data and design method-ologies.Through knowledge reasoning,the graph structure is refined to establish associations between road segments and design methodologies.The system implements natural language query capabilities through term mapping,large language model(LLM)parameter parsing,and local query templates,and fea-tures a visual interface that intuitively displays knowledge query results.Application analysis indicates that the system can perform road segment profiling,identify shortcomings,trace the root causes of problems,recommend design methods,and automate the entire"evaluation– diagnosis– remediation"process.Re-search indicates that the knowledge graph system can organize roads,indicators,issues,and design meth-ods into an interpretable logical structure,providing effective support for the refined management of side-walks and non-motorized vehicle lanes.
刊 名 城市交通 ISTIC
年,卷(期): 2026, 24(5)
分类号: U491.1+2
机标分类号: TP391 U491 TP29
在线出版日期: 2026年11月2日
基金项目: 2023YFC3805300:国家重点研发计划,