基于多模式交通网络建模的居民出行特征探测——以深圳市为例  

Detection of Resident Travel Characteristics Based on Multi-modal Transportation Network Modeling:A Case Study of Shenzhen

作　　者：杨梦勤 王艳慧[1,2] 罗霄月 蔡鑫雨 刘佳皓 YANG Mengqin;WANG Yanhui;LUO Xiaoyue;CAI Xinyu;LIU Jiahao(College of Resource Environment and Tourism,Capital Normal University,Beijing 100048;Key Laboratory of 3DInformation Acquisition and Application,Ministry of Education,Capital Normal University,Beijing 100048,China)

机构地区：[1]首都师范大学资源环境与旅游学院,北京100048 [2]首都师范大学三维信息获取与应用教育部重点实验室,北京100048

出　　处：《地理与地理信息科学》2023年第6期1-8,17,共9页Geography and Geo-Information Science

基　　金：国家自然科学基金项目(42171224);国家重点研发计划项目(2018YFB0505400);北京市长城学者资助项目(CIT&TCD20190328)。

摘　　要：当前研究多基于单一数据源模拟单一出行模式的交通网络,难以客观全面地反映居民真实出行特征。该文以深圳市为例,提出基于交通小区划分和站点路径匹配的多模式交通网络建模方法和居民出行特征分析的多特征融合算法。首先,顾及交通网络时空分异性,设计多尺度交通小区划分算法简化复杂网络;然后,利用站点路径属性与交通小区匹配连接,构建多模式集成的复杂交通网络模型;最后,综合利用移动轨迹挖掘和特征融合技术,设计多模式出行特征融合算法,系统刻画多模式融合下的居民出行特征。实验结果表明:相比仅使用单一数据源,多模式集成方法能有效消除数据偏差,更准确地揭示区域间的交通紧密程度,更全面地反映居民的出行行为和偏好,可以更好地为城市规划和交通治理提供科学决策依据。Current research predominantly relies on single travel data sources to simulate single-modal transportation networks,making it challenging to objectively and comprehensively reflect the real travel characteristics of residents.This paper presents a modeling method of multi-modal transportation network and a multi-feature fusion algorithm for analyzing resident travel characteristics,taking Shenzhen as a case study.Firstly,taking into consideration the spatiotemporal variations in transportation networks,a partitioning algorithm of multi-scale transportation zones is designed to simplify the complex network.Next,a multi-modal transportation network is constructed by connecting station path attributes with transportation zones.Finally,mobile trajectory mining and feature fusion techniques are adopted to design a fusion algorithm of multi-modal travel features,systematically depicting resident travel characteristics in a multi-modal context.Experimental results demonstrate that compared to using a single data source alone,the proposed method effectively mitigates data biases,provides a more accurate representation of traffic density among different transportation regions,and offers a more comprehensive insight into residents′travel behaviors and preferences,and can better provide scientific decision-making for urban planning and transportation management.

关 键 词：多模式交通网络建模 交通小区 特征融合 出行特征 

分 类 号：U491.112[交通运输工程—交通运输规划与管理] P208[交通运输工程—道路与铁道工程]