基于城市用地的公共自行车骑行量建模  

Public bicycle usage modeling based on urban land use

作　　者：唐炉亮[1] 陈西 杨雪[1] 张霞[2] 李清泉[1,3] TANG Lu-liang;CHEN Xi;YANG Xue;ZHANG Xia;LI Qing-quan(State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan 430079, Hubei, China;School of Urban Design, Wuhan University, Wuhan 430070, Hubei, China;College of Civil Engineering, Shenzhen University, Shenzhen 518060, Guangdong, China)

机构地区：[1]武汉大学测绘遥感信息工程国家重点实验室,湖北武汉430079 [2]武汉大学城市设计学院,湖北武汉430070 [3]深圳大学土木工程学院,广东深圳518060

出　　处：《交通运输工程学报》2018年第1期150-158,共9页Journal of Traffic and Transportation Engineering

基　　金：国家重点研发计划(2017YFB0503604;2016YFE0200400);国家自然科学基金项目(41271442;41571430;41671442);教育部联合基金项目(6141A02022341)

摘　　要：将公共自行车出行模式划分为前往站点、租车骑行和还车离开3个阶段,分别研究了城市用地在每个阶段对公共自行车骑行量的影响;针对第1、3阶段,采用兴趣点度量公共自行车站点周围的城市用地类型,通过引入时间满意度函数,基于多元回归分析,提出了预测骑行产生量和吸引量的建模方法;针对第2阶段,构建了城市用地对公共自行车骑行量的6个影响指标,包括起始站点骑行产生量、起始站点分布密度、起始站点偏离度、终止站点骑行吸引量、起终站点间距与起终站点用地差异;对纽约市961 865条以及上海市1 185 816条公共自行车骑行记录进行多元回归分析。分析结果表明:在保持用地因素数量不变的情况下,传统方法针对纽约和上海记录数据的决定系数分别为0.581、0.474,建模方法的决定系数较高,分别达到了0.738、0.607;针对自行车站点间客流量的建模,提出的6个指标均对站点间客流量有显著影响,并且模型的调整决定系数为0.487。可见,建模方法更加合理地衡量了城市用地对公共自行车骑行量的影响,并对站点的骑行产生量、站点间客流量与站点吸引量实现了更为准确的建模。The public bicycle travel pattern was divided into three phases, namely going to a station, riding after renting, and leaving after returning. The influence of urban land use on public bicycle usage was investigated for each phase. For the first and third phases, the points of interst were used to measure the type of urban land use surrounding public bicycle station. By introducing time satisfaction function, a modeling method of the prediction of public bicycle demand and attraction was proposed based on multiple regression analysis. For the second phase, six influence factors of public bicycle usage related to urban land use were established, including the bicycle demand of origin station, the density of origin station, the remoteness of origin station, the attraction of destination station, the distance between origin and destination station, and the land use difference between origin and destination station. 961 865 and 1 185 816 records of public bicycle usage respectively from New York and Shanghai were analyzed by using multiple regression analysis. Analysis result indicates modeling method yields higher determination that for the records of New York and Shanghai, the coefficients（0. 738 and 0. 607） than the traditional method （0. 581 and 0. 474） when the number of land use factors is invariable. All of six proposed factors have significant influence on the traffic volumes between the stations in bicycle traffic volume model, and the adjusted determination coefficient is 0. 487. In conclusion, the modeling method can measure the influence of urban land use on the public bicycle usage more feasibly, and it provides a more accurate model of trip demand at stations, traffic volume between stations, and attraction of stations for public bicycle trips. 5 tabs, 7 figs, 25 refs.

关 键 词：交通信息工程 公共自行车骑行量 时间满意度函数 城市用地 出行模式 多元回归 

分 类 号：U491.1[交通运输工程—交通运输规划与管理]