基于出租车轨迹数据的载客情况可视化分析

发布时间：2018-01-06 05:42

本文关键词：基于出租车轨迹数据的载客情况可视化分析　出处：《浙江工业大学》2014年硕士论文　论文类型：学位论文

【摘要】：城市化促进了社会经济的发展,改善了人们的生活水平,同时也带来了诸如道路拥堵、出行耗时长等交通问题。为了能够了解出行情况,缓解交通问题,交通管理部门将越来越多诸如GPS的设备安装在出租车上。通过这些设备的信息采集,形成大量的出租车轨迹数据。出租车轨迹数据隐含了大量知识能够帮助我们分析人们出行信息,达到优化交通、改善路况的目的。然而出租车轨迹数据本身量大且相对复杂,仅仅依靠数据本身很难让人直观了解,可视分析技术为我们提供了一种有效展示和分析数据的方法。本文中我们利用出租车轨迹数据来分析人们出行的上下客热点,帮助出租车司机结合道路交通情况找出较优的载客行驶轨迹,同时开发一个可视化原型系统来支持可视化分析。本文的研究内容主要包括以下几个方面：(1)出租车轨迹数据预处理。通过设定轨迹提取的规则和方法,对轨迹数据中存在的跳变点轨迹、停车轨迹、过短轨迹进行过滤,提取得到出租车轨迹。(2)上下客热点提取和分析。根据出租车轨迹提取得到上下客点,在此基础之上设计了一种改进的上下客热点生成聚类算法-—GBADBSCAN来生成上下客热点。采用基于上下客热点聚类图标方法对不同时段上下客热点的分布进行可视化分析。(3)出租车推荐行驶轨迹的获取。本文首先通过基于起终点的相似轨迹方法来将所有轨迹划分成具有相近起点和终点的出租车轨迹子集,接着采用基于密度的ε距离轨迹聚类算法来对轨迹子集聚类,找出不同上下客热点间的候选行驶轨迹。最后结合候选轨迹的行驶时间、速度、距离以及载到乘客的可能性来设置权值,根据带权轨迹树来寻找空载出租车到就近上客热点载客的最优行驶轨迹。(4)构建可视化原型系统。本文开发了地图、时间、平行坐标、控制台以及路径导航描述组件这5种组件来构建可视化原型系统,实现出租车轨迹数据的可视化分析。
[Abstract]:Urbanization has promoted the development of social economy, improved people's living standards, but also brought traffic problems such as road congestion, long travel time, in order to understand the travel situation and alleviate the traffic problems. More and more devices, such as GPS, are being installed in taxis by traffic authorities. A large number of taxi trajectory data are formed. Taxi trajectory data implied a lot of knowledge can help us to analyze people travel information to achieve traffic optimization. The purpose of improving road conditions. However, taxi track data itself is large and relatively complex, relying on the data itself is difficult to intuitively understand. Visual analysis technology provides us with an effective method to display and analyze data. In this paper, we use taxi trajectory data to analyze the hot spots of people travelling. Help taxi drivers to find out the best path to carry passengers in combination with road traffic conditions. At the same time, a visual prototype system is developed to support visual analysis. The research content of this paper mainly includes the following aspects: 1) the pretreatment of taxi trajectory data. By setting the rules and methods of trajectory extraction. The jump point locus, parking track, too short track in the track data are filtered, the taxi track is extracted and the hot spot is extracted and analyzed. According to the taxi trajectory extraction, the upper and lower passenger points are obtained. On this basis, an improved clustering algorithm named GBADBSCAN is designed to generate hot spots. For visual analysis. 3) the acquisition of taxi recommended trajectory. Firstly, this paper divides all the tracks into subsets of taxi tracks with similar starting point and end point by using the similar trajectory method based on the starting end point. Then the 蔚 distance trajectory clustering algorithm based on density is used to cluster the trajectory subset to find out the candidate trajectory between different hot spots. Finally, combining the travel time and speed of the candidate trajectory. Distance and the possibility of carrying a passenger to set the weight value. Based on the weighted track tree to find the optimal driving path of no-load taxi to the nearest hot spot, a visual prototype system is constructed. In this paper, map, time, parallel coordinates are developed. The 5 components of the console and the path navigation description component are used to construct the visual prototype system to realize the visual analysis of the taxi track data.
【学位授予单位】：浙江工业大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495;TP311.13

【相似文献】