基于手机信令数据和Needleman-Wunsch序列匹配算法的都市圈出行方式识别

Metropolitan Travel Mode Recognition Based on Mobile Phone Data and Needleman-Wunsch Sequence Matching Algorithm

为弥补传统的居民出行调查方法在都市圈出行调查中的不足,提出一套基于手机信令数据的都市圈跨市出行方式识别算法,通过分析不同出行方式在出行路径、出行时间、出行地点方面的差异,实现高铁、大巴及小汽车3类主要跨市出行方式识别。首先,在成都市成华区和资阳市雁江区设计都市圈出行试验,招募志愿者采集出行手机信令数据;然后,分析都市圈跨市出行连接信号基站分布特征,再基于Needleman-Wunsch序列匹配算法提取出行路径信息,区分高铁出行和公路出行;最后,针对公路出行方式构建模糊识别模型,划分小汽车和大巴车出行。研究结果显示,在志愿者出行识别中,通过与志愿者出行日志对比,所提模型出行路径识别准确率达93.80%,3种出行方式均被正确识别;在研究区域全部手机用户集计出行识别中,3种出行方式占比大致为12.79%,1.36%,85.85%,与成都交通发展研究院2021年公布的《成德眉资区域出行报告》中出行分担比基本相同(12.23%,1.36%,86.41%),证明所提算法识别效果良好。

To make up for the deficiencies of traditional resident travel survey methods in metropolitan area travel survey,a set of metropolitan cross-city travel mode identification algorithms based on mobile phone signaling data were proposed.Three main modes of cross-city travel including highspeed rail,bus and car,were identified by analyzing their differences in travel paths,time,and location.Firstly,the metropolitan travel experiment between Chenghua District of Chengdu and Yanjiang District of Ziyang was designed,and volunteers were recruited to collect their mobile phone signaling data in actual metropolitan cross-city travel.Secondly,the distribution characteristics of cross-city travel connection signal base stations in metropolitan areas were analyzed,and the travel routes were extracted to distinguish between high-speed rail travel and highway travel based on the Needleman-Wunsch sequence matching algorithm.Finally,the fuzzy recognition model for road travel mode was constructed to divide the car and bus travel.The research results show that in volunteer travel identification,the travel path identification accuracy of the proposed model was up to 93.80%compared with volunteer travel logs,and all three travel modes were correctly identified.While in the aggregate travel identification of all mobile phone users in the study area,the rates of the three travel modes were12.79%,1.36%,and 85.85%,which was basically the same as the travel share ratio in the ChengduDeyang-Meishan-Ziyang Regional Travel Reports published by Chengdu Transportation Development Research Institute in 2021(12.23%,1.36%,86.41%).It is proved that the proposed algorithms have good recognition effect.