基于智能手机的个体活动链信息获取及挖掘方法

项目名称： 基于智能手机的个体活动链信息获取及挖掘方法

项目编号： No.51478266

项目类型： 面上项目

立项/批准年度： 2015

项目学科： 建筑科学

项目作者： 隽志才

作者单位： 上海交通大学

项目金额： 80万元

中文摘要： 本项目以时空轨迹为数据源，提出个体活动-出行信息挖掘方法，设计不完整活动链弥补方法，研究个体、家庭和小区属性对多日活动-出行行为的影响机理。主要内容包括：基于智能手机和网络调查技术采集个体出行轨迹，运用时空路径和时空棱镜模型，实现轨迹在时空维度下的可视化；运用卡尔曼滤波地图匹配算法，实现轨迹与地图的有效匹配，结合GIS信息和POI知识数据库，分别使用基于特定规则的算法、离散隐马尔可夫模型和时空匹配方法推断出行端点、方式和目的；提出活动链完整性边界条件，建立不完整活动链的分类体系，设计每种不完整活动链特有的弥补规则；生成个体、家庭和小区三个层级的活动-出行影响因素集合，建立以集合中影响要素为变量的多日活动-出行模型，精细化分析个体多日活动-出行特性和规律。项目研究成果在理论上能够丰富和扩展出行行为研究的数据获取方法和理论框架，在实践上为精细化分析交通需求特征和制定可行的交通管理政策提供依据。

中文关键词： 出行行为；个体活动链；数据挖掘；智能手机；交通需求特征

英文摘要： This project studies the influence mechanism of individual, household and analysis zone characteristics on multiday activity-travel behavior using individual temporal and spatial tracks as main data source. It aims to develop operational methods of incomplete activity chain imputation and activity chain characteristic data mining.The main contents of this project are as follows. First, travel trajectory data will be acquired from smartphone-based travel survey and web-based survey, and then be visualized in space and time by using space-time paths and prisms. Then, algorithms will be provided to match the trajectory data stream to coded map of the transportation network with the application of Kalman filtering. Based on this, trip ends, travel mode, and purpose will be imputed from a combination of GIS and POI data using rule-based methods, discrete hidden Markov models, and space-time matching algorithms. Next, the boundary condition of complete activity chains is determined, and the classification standards of incomplete activity chains are established，and imputation rules for each class are developed. Finally, activity-travel behavior influence factors are analyzed at a hierarchy of individual, household and analysis zone, and a multi-day model analyzing the effect of the variables in the set on the activity-travel behaviors is developed, and explicit analysis of multi-day activity-travel characteristics and rule is implemented. Results from this research will expand and enrich the theoretical framework and data collection methods of travel behavior analysis in theory, and will help to explicitly analyze activity-travel characteristics and rule and decide traffic management policies.

英文关键词： Travel behavior;Individual activity chain;Data mining;Smartphone;Transportation demand characteristics