【作者】 姜凡；

【导师】 赵东；

【作者基本信息】 北京邮电大学 ， 计算机技术（专业学位）， 2021， 硕士

【摘要】 电动出租车轨迹描述了电动出租车空间位置和其他属性随时间的变化,蕴含了大量的居民出行信息,也反映了电动出租司机的充电习惯和出行特征。本文通过轨迹数据可视分析方法对深圳电动出租车的轨迹数据进行研究,并开发出电动出租车轨迹数据可视分析系统,支持多维度、全方面分析电动出租车在深圳的推广运营情况,从而对电动出租车行业的健康发展提出建议。本文的研究内容主要包括以下几个方面:1.充电事件提取及可视分析。根据电动出租车司机的充电行为设计充电事件提取算法,并对提取出来的充电事件进行不同维度的统计分析。通过对充电事件的可视分析,帮助电动出租车司机合理规划充电时间,选择合适的充电站减少充电需要花费的等待时间。2.上下客热区提取及可视分析。首先从电动出租车的轨迹数据中提取出不同时段乘客的上下客点,然后通过参数自适应的聚类算法(DBSCAN)对提取出来的上下客点进行了聚类,生成上/下客热区。通过对上下客热区的可视分析,对司机寻客区域提出建议,减少司机不必要的寻客时间,有效地帮助司机提高收入。3.电动出租车轨迹可视分析系统开发。首先根据特定的清洗规则对电动出租车轨迹数据进行清洗,提高分析数据的质量。然后针对关注的重点特征进行提取,最后开发了电动出租车轨迹数据可视分析系统,将这些特征用多图表联动的形式展示出来,为多维度、全方面分析电动出租车的推广运营状况提供了有力支持。更多还原

【Abstract】 The trajectory of electric taxis describes the change of spatial location and other attributes of electric taxis with time,contains a large amount of residents’ travel information,and also reflects the charging habits and travel characteristics of electric taxi drivers.This paper studies the trajectory data of Shenzhen electric taxis through the visual analysis method of trajectory data,and develops the visual analysis system of trajectory data of electric taxis,which supports multi-dimensional and comprehensive analysis of the promotion and operation of electric taxis in Shenzhen,so as to put forward suggestions for the healthy development of electric taxi industry.This paper mainly includes the following aspects:1.Charging event extraction and visual analysis.The charging event extraction algorithm is designed according to the charging behavior of electric taxi drivers,and the extracted charging events are statistically analyzed in different dimensions to visually display the charging selection preferences of different drivers and the charging status of the charging station.Through visual analysis of charging events,we can help electric taxi drivers to reasonably plan the charging time and choose the appropriate charging station to reduce the waiting time required for charging.2.Passenger hot spots extraction and visual analysis.Firstly,the pick-up and drop-off points of passengers in different periods are extracted from the trajectory data of electric taxis,and then the pick-up and drop-off points are clustered by the parameter adaptive clustering algorithm(DBSCAN)to generate the pick-up and drop off hot spots.Through the visual analysis of the hot area,it can reduce the unnecessary time for drivers to find customers and effectively help drivers to improve their income.3.Development of visual analysis system for track of electric taxis.Firstly,the trajectory data of electric taxis is cleaned according to specific cleaning rules to improve the quality of analysis data,then the key features are extracted,finally develop a visual analysis system for electric taxis’trajectory data.These features are displayed in the form of multi-chart linkage,which provides strong support for multi-dimensional and comprehensive analysis of the promotion and operation of electric taxis.更多还原