基于数据分析的城市移动模式挖掘

发布时间：2018-05-12 10:33

本文选题：地图匹配 + 交通流　；参考：《东南大学》2015年硕士论文

【摘要】：智慧城市是充分利用城市各行各业的数据,采用信息技术综合处理分析,为城市各行各业提供智能服务,而城市移动模式是智慧城市的基础性课题。本文基于南京七千多辆出租车两个月的GPS数据,采用数据分析的方法对南京城区移动对象的移动模式进行了研究。文章主要在地图匹配(Map Matching, MM)、城市车流模式、城市人群移动模式三个方面展开工作。城市车流模式聚焦在日常交通流量和常发拥堵路段的时空分布,而城市人群移动模式则重点关注城市人群出行的距离分布、人群出行热点以及人群流向的时空变化。文中本地化的结论为城市交通规划、区域规划和公共卫生建设提供了指导性知识。文中首先综述了城市移动模式的研究背景和研究现状,然后简介了开放街道地图(Open Street Map, OSM),并详细分析了出租车GPS数据的噪声。过滤噪声之后,本文使用局部ST-Matching算法将出租车GPS数据匹配到OSM电子地图上。之后对南京不同等级道路在工作日和非工作日的交通流量和行车速度进行统计分析,发现单一道路等级在不同时间段上行车速度服从正态分布。本文以此为基础,建立路段拥堵分值模型,引入常发拥堵指标之后完成对南京常发拥堵点的提取。发现市中心的中山南路、内环线、二桥南路和大桥高速等路段是城市路网的瓶颈,经常发生拥堵。其次,本文对南京城市人群打车出行的距离和不同时间段打车出行的人数进行统计分析,验证了打车出行,人群移动距离服从幂率的结论。并使用基于模块度的层次聚类算法对南京城市人群出行热点进行分析,发现在早晚高峰人群的出行热点主要集中在中华门、安德门、迈皋桥地铁站等主要的交通换乘点和大型住宅区,而南京火车站和江苏省人民医院等大型医院在一天的大多数时间内都是人们出行的热点,凌晨时分1912街区是人们主要的出行热点。最后,本文采用栅格统计的方法,对城市人流动向进行分析,发现在不同时段人群流向表现出很大不同,早高峰主要是住宅区到车站,晚高峰则刚好相反,表现出较好的对称性。在其他时间段,则语义呈现多样性。
[Abstract]:The intelligent city is to make full use of the data of various industries in the city, and to adopt the information technology to deal with the analysis synthetically, to provide the intelligent service for the various industries of the city, and the urban mobile mode is the basic subject of the intelligent city. Based on the GPS data of more than 7,000 taxis in Nanjing for two months, this paper studies the moving pattern of moving objects in Nanjing urban area by using the method of data analysis. This paper mainly focuses on map matching, MMX, urban traffic flow and urban crowd movement. The urban traffic flow mode focuses on the spatial and temporal distribution of daily traffic flow and common congested sections, while the urban crowd movement mode focuses on the distance distribution of urban population travel, crowd travel hot spots and the spatio-temporal change of crowd flow direction. The conclusion of localization provides guiding knowledge for urban transportation planning, regional planning and public health construction. In this paper, the research background and present situation of urban mobile mode are summarized, then the open Street map is introduced, and the noise of taxi GPS data is analyzed in detail. After filtering noise, this paper uses local ST-Matching algorithm to match taxi GPS data to OSM electronic map. Then the traffic flow and driving speed of different grades of roads in Nanjing on weekdays and non-workdays are statistically analyzed and it is found that the single grade of roads in different periods of time from normal distribution of driving speed. Based on this model, the model of traffic congestion score is established, and the normal congestion index is introduced to complete the extraction of normal congestion points in Nanjing. It is found that Zhongshan South Road, Inner Ring Road, second Bridge South Road and Bridge Expressway are the bottleneck of urban road network, and congestion often occurs. Secondly, this paper makes a statistical analysis on the distance of the urban population in Nanjing and the number of people traveling in different time periods, and verifies the conclusion that the power ratio of the driving distance of the crowd is the same as that of the ride-hailing trip and the crowd moving distance. And using the hierarchical clustering algorithm based on modular degree to analyze the travel hot spots of Nanjing city crowd, it is found that the hot spots in the morning and evening rush crowd mainly focus on the Zhonghua Gate, the Anderman Gate. Megaoqiao subway stations and other major transportation and residential areas, while Nanjing Railway Station and Jiangsu Provincial people's Hospital and other large hospitals during most of the day is a hot spot for people to travel. The 1912 block in the early hours of the morning is the main hot spot for people to travel. Finally, by using the method of grid statistics, this paper analyzes the trend of urban passenger flow. It is found that the flow of people in different periods is very different. The early rush hour is mainly from residential area to the station, while the late rush hour is just the opposite. Show good symmetry. In other time periods, semantic diversity is present.
【学位授予单位】：东南大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TU984.191;TP311.13

【参考文献】