基于手机数据的城市交通路网预测与路径选取

发布时间：2018-06-02 21:30

本文选题：交通预测 + 路径选取　；参考：《北京交通大学》2014年硕士论文

【摘要】：摘要：智能交通系统(Intelligent Transportation Systems,ITS)是目前国际公认的缓解城市交通拥堵、减少交通事故和降低汽车尾气排放污染等交通问题的最佳途径。道路交通信息是ITS不可或缺的前提和主要内容。在短时间得到准确的交通信息,并根据这些信息快速准确地为驾驶员提供最佳行驶路径,是ITS领域的一个重要研究方向。目前,随着智能手机的普及,它在给人们带来通信便利的同时还形成了大量的交通数据。通过对这些手机交通数据的采集和处理,用于交通路网的短时预测及路径诱导已成为ITS的一个前沿课题。本文根据现有条件,主要研究利用手机数据进行交通路网预测和路径选择的建模和算法实现问题,主要内容如下：第一,根据交通路网特性和手机交通数据特点,引入虚拟路径(VTLs)的思想,设计了基于交通路网的时空自回归移动平均模型(STARMA)和逻辑回归算法(LR),分别用于预测交通路网的旅行时间和拥堵状况。第二,根据时空自回归移动平均模型和逻辑回归算法,提出一个基于连通可靠性的路径选取算法。通过逻辑回归算法得到的拥堵概率数据作为诱导路径中的旅行时间权重值,使得在路径选取过程中更倾向于那些没有或很少拥堵的路径。从而解决了一般算法中一味地追求最少的时间花费反而导致局部地区车辆的激增,进而影响交通路网可靠性的问题。第三,依托Paramics仿真软件通过编程,实现了对利用手机采集交通数据过程的模拟。利用Paramics仿真软件对路网进行建模及仿真,并将STARMA和LR的仿真结果与真实数据进行对比分析；另外,将上述路径选取算法的仿真结果与其他组合算法进行对比分析。
[Abstract]:Abstract: intelligent Transportation (Intelligent Transportation) is the best way to alleviate traffic congestion, reduce traffic accidents and reduce vehicle exhaust pollution. Road traffic information is an indispensable prerequisite and main content of ITS. It is an important research direction in the field of ITS to get accurate traffic information in a short time and provide drivers with the best driving path quickly and accurately according to these information. At present, with the popularity of smart phones, it not only brings convenience to people, but also forms a lot of traffic data. Through the collection and processing of these mobile phone traffic data, short-term prediction and route guidance of traffic network has become a frontier topic of ITS. According to the existing conditions, this paper mainly studies the modeling and algorithm implementation of traffic network prediction and path selection using mobile phone data. The main contents are as follows: First, according to the characteristics of traffic network and mobile phone traffic data, the idea of virtual path (VTLs) is introduced. A spatio-temporal autoregressive moving average model based on traffic network (STARMA) and a logical regression algorithm (LRN) are designed to predict the travel time and congestion of traffic network respectively. Secondly, a path selection algorithm based on connected reliability is proposed based on the moving average model and logical regression algorithm. The congestion probability data obtained by the logical regression algorithm is used as the travel time weight value in the induced path, which makes the path selection more prone to those paths that have no or little congestion. Thus solving the problem that the pursuit of the least amount of time in the general algorithm leads to the proliferation of vehicles in local areas and thus affects the reliability of the traffic network. Thirdly, the simulation of traffic data acquisition using mobile phone is realized by programming with Paramics simulation software. The simulation results of STARMA and LR are compared with real data by using Paramics simulation software. In addition, the simulation results of the above path selection algorithm are compared and analyzed with other combined algorithms.
【学位授予单位】：北京交通大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U495

【参考文献】