基于公交GPS和IC卡数据的公交OD推算

发布时间：2018-03-24 15:04

本文选题：数据挖掘　切入点：公交IC卡数据　出处：《哈尔滨工业大学》2017年硕士论文

【摘要】：城市居民公交出行OD是城市公共交通规划的基础资料,是城市公交线网布局和优化的重要依据。传统居民公交出行OD采集采用的是人工调查方法,不仅耗费大量人力、物力,还存在采集的数据样本量少、精度低等问题。随着公交刷卡收费系统和GPS系统等信息技术被逐渐引入城市公交系统中,随之产生了大量的公交IC卡数据,这为城市公共交通基础数据采集开辟了新的途径。但目前国内多数城市公交运营仍采用一票制模式,公交IC卡数据只有上车记录不包含下车记录,无法直接获取居民出行信息。因此,如何有效地从这些数据中获取居民公交出行OD,对城市公共交通服务水平提升有着重要的现实意义。为实现这一目标,本文将对公交IC数据和GPS数据进行分析,以推算出居民出行OD矩阵。本文首先对公交IC卡和GPS数据进行分析处理,根据公交IC卡数据特点,提出了一种推算乘客刷卡比例的方法。针对一票制公交IC卡缺少乘客站点信息的特点,将公交GPS记录时间与公交IC卡刷卡时间匹配,获取乘客上车时位置信息,识别乘客上车站点;采用时空约束进行乘客换乘识别,以乘客相邻两次刷卡记录时间间隔和两个站点间的空间距离为识别标准进行换乘的识别。在下车站点识别中,在参考深圳居民出行特点的同时,本文提出引入地铁IC卡数据作为补充数据,完善乘客出行链,采用基于出行链的乘客下车站点概率模型进行下车站点的推算。在完成上下车站点识别基础之上,建立公交站间OD矩阵,以刷卡比例与识别成功率进行扩样。根据站点与交通小区的空间关系,将公交站间OD矩阵转化为居民公交出行OD矩阵。最后以深圳市为例进行居民出行OD矩阵推导,验证算法的有效性和可行性。
[Abstract]:Urban residents travel OD is the basic data of urban public transport planning, it is an important basis for the layout and optimization of urban transit network. The traditional OD collection of urban public transport travel is a manual survey method, which not only consumes a lot of manpower and material resources. There are still some problems such as small sample size and low precision. With the introduction of information technology, such as charging system and GPS system, a lot of IC card data are produced. This opens up a new way to collect the basic data of urban public transport. However, at present, most cities in China still use a one-ticket system in their public transport operations. The IC card data of public transportation only contains no record of getting off or off the bus. Therefore, how to obtain the ODO from these data effectively has important practical significance to improve the level of urban public transport service. In order to achieve this goal, In this paper, the IC data and GPS data of public transportation are analyzed to calculate the OD matrix of residents travel. Firstly, the data of IC card and GPS are analyzed and processed, according to the characteristics of IC card data, This paper presents a method of calculating the proportion of passengers swiping cards. In view of the lack of station information on IC cards of one ticket system, the GPS recording time of public transportation is matched with the time of swiping cards of public transportation IC cards, and the location information of passengers when boarding on board is obtained. Identification of passenger boarding station; use of temporal and spatial constraints to identify passenger transfer, using two adjacent swipe card recording time interval and space distance between the two stations as the identification criteria. While referring to the travel characteristics of residents in Shenzhen, this paper proposes to introduce IC card data as supplementary data to improve the passenger travel chain. The probability model of passenger and disembarkation station based on trip chain is used to calculate the station. On the basis of the identification of boarding and disembarking station, the OD matrix between bus stations is established. According to the spatial relationship between the station and the traffic district, the OD matrix between the bus stations is transformed into the OD matrix of the residents' public transport trip. Finally, taking Shenzhen City as an example, the OD matrix of the residents' trip is deduced. The validity and feasibility of the algorithm are verified.
【学位授予单位】：哈尔滨工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491.17

【参考文献】