基于手机数据的城市道路车源预测理论研究

面向当代严重的城市交通拥堵，城市交通规划与组织水平受到人们的广泛关注，交通需求作为交通规划与组织的重要基础数据，其获取方法一直以来都是非常重要的研究课题。由于传统的交通需求预测方法耗费巨大的人力物力，利用各种信息资源进行数据挖掘，分析获取交通需求成为了新的研究方向。基于大规模手机数据和人类出行建模领域的最新成果，本项目研究城市道路车源预测理论。主要研究内容包括：基于手机数据的城市居民出行建模、基于手机数据的交通需求预测、城市道路车源预测、导致拥堵的车源分析。通过运用手机数据准确预测居民实时位置，计算居民出行分布、估计O-D矩阵，找到一种经济、可行的交通需求预测方法。本项目将人类出行建模领域的最新成果引入到交通需求预测中，通过产生道路车流的人的实时位置掌握车流的形成过程，动态分析各路段的车辆来源，以期搭建人类出行行为研究领域与交通需求预测研究领域的桥梁。

Currently, we are faced with severe urban traffic congestions. The urban traffic planning and organizing level has been paid much attention. As an important fundamental data for ubran traffic planning and organization, traffic demand has been an important research topic for a long time. Since it is expensive to predict traffic demand by traditional surveys, as a new reseach direction, many new information sources are used. Based on large scale mobile phone data, we propose the theory of driver source prediction. The main contents of this project include: resident mobility model based on mobile phone data, prediction of traffic demand based on mobile phone data, prediction of driver sources on urban roads, analysis of the driver sources that cause traffic congestions. Using mobile phone data, we aim to accurately predict the residents' hourly locations and calculate trip distribution and O-D matrix. Through this method we try to find an economic and feasible way to predict traffic demand. By introducing the newest results of human mobility modeling to traffic demand prediction, we aim to understand the process of the formation of traffic flow by finding drivers' hourly locations. We will dynamically analyze the driver sources on roads, aiming to build a bridge between the fields of human mobility modeling and traffic demand prediction.