多源轨迹数据驱动的城市居民交通出行活动链特征模式研究

In the paradigm of geographical big data, the increasing individual spatio-temporal trajectory data provide new perspectives to conduct individual-based studies on urban traffic analysis issues. Most existing human mobility literature could not get the ideal results for the traffic prediction due to the logistical restraints of recording activity information and sociology analysis. By adopting the traffic bayonet monitor data and social media check-in data, this research project aims to: (1) To explore the sematic trajectory, we extract the trip-activity chain based on the mobility pattern and the spatio-temporal structures of intra-urban. (2) To explain the mechanism of sociology of the residents travel mode, we explore the Trip-Activity Chain Character Pattern (TACCP) by taking account of sociology factors. (3) We then proposed an intra-urban traffic motion model based on the TACCP and depicted the evolvement characteristic of network traffic flow. The implications of this research project can benefit a better understanding the interaction between the micro individual activity pattern and the macroscopic urban road traffic flow, and provide a theoretical support for traffic route guidance and safety control.

地理大数据研究范式下，与日俱增的个体时空轨迹数据为从微观个体移动特征进行宏观城市交通问题分析提供了新的研究思路。然而目前大部分研究由于缺乏大规模、长时间个体活动链数据及分析手段支持，并且忽略了活动链背后的社会学因素分析，无法对道路交通需求分布进行准确模拟和预测。本项目基于交通卡口和位置签到等多源轨迹数据，（1）提出了基于时空约束和城市时空结构的出行活动链提取方法，对出行轨迹进行语义信息匹配。（2）在此基础上，提出了顾及人文社会因素的出行活动链特征模式挖掘方法，将个体活动模式与人文社会因素进行关联分析，以解释居民出行模式的社会学驱动机制。（3）最后，提出了基于出行活动特征模式的城市居民交通移动模型，对道路交通流的时空演化过程进行模拟和预测。本研究将有助于更好地理解微观个体活动模式对于宏观城市道路交通流形成与变化的影响，将为道路交通的微观诱导和安全管控提供科学支撑。