文章编号： 1672-5328（2016）01-0009-06

杨飞，姚振兴
(西南交通大学交通运输与物流学院，四川成都610031)

摘要： 手机调查方法的已有研究较多集中于基于手机信令数据的宏观出行特征获取，而手机传感器 数据在个体出行链微观出行特征提取方面具有优势。针对城市居民多采用组合交通方式出行的特 征，研发智能手机应用软件，实现GPS数据(位置坐标与速度)、加速度计、服务基站、WiFi 等传感 器数据采集。运用小波分析、神经网络等数据挖掘技术分析不同交通方式出行数据差异，探索多种 数据挖掘算法用于个体出行参数提取的可行性及效果。结合实际案例，总结应用手机传感器数据进 行出行特征精细化提取的难点和技术关键。最后，探讨精细化个体出行数据在交通模型和理论优化 方面的应用。

关键词： 大数据；智能手机；传感器数据；出行特征；数据挖掘；交通模型优化

中图分类号： U491

文献标识码：A

Cellular-based Data Extracting Method for Travel Characteristics

Yang Fei, Yao Zhenxing
(School of Transportation and Logistics, Southwest Jiaotong University, Chengdu Sichuan 610031, China)

Abstract： Existing research of cellular-based survey methods mainly focus on travel characteristics at macro level. It should be known that cellular probe data also have great advantages of extracting travel characteristics at micro level – individual travel chains. Considering majority of urban residents' multimodal travel patterns, a mobile app is developed to retrieve traveler disaggregated data, such as GPS (coordinates and speed), accelerometer through base station and Wi-Fi connectivity. This paper analyzes the difference in data from various travel modes using wavelet analysis, neural network and other data mining techniques. The feasibility and multiple data mining algorithms used to extract individual travel parameters are discussed. Based on case studies, the paper summarizes difficulties and key technical points of using cellular probe data to extract accurate travel characteristics. Finally, the paper discusses the application of individual travel data in transportation modeling.

Keywords： big data; smartphone; probe data; travel characteristics; data mining; optimization of transportation models