【作者】 李强；

【导师】 孙敏； 庞溟舟；

【作者基本信息】 上海交通大学 ， 控制科学与工程， 2016， 硕士

【摘要】 随着汽车普及程度的提高,城市中的拥堵现象越来越严重,出行路径选择越来越受到人们的关注。近年来,一方面,网络技术发展迅速,使得各方面的数据的获取要容易很多;另一方面,数据挖掘技术的兴起带来了数据研究的热潮。基于以上阐述,本文利用出租车的轨迹数据进行城市交通小区之间的路径选择问题的相关研究,取得了如下研究成果:1)基于上海市出租车轨迹数据的交通小区之间路径规划算法实现。主要针对选取的出租车轨迹数据进行挖掘,从而提供一种新的思路来研究路径选择问题。通过与谱聚类结合,很好地解决了轨迹数据在维度上的不平衡性。分析并改进了当前谱聚类算法的不足(针对高斯核尺度参数进行优化)。最后提出了局部层级路网的概念,从局部研究复杂路网中的路径选择问题。实验表明所提出的路径分类算法以及路径选择算法具有一定的可执行性。2)上海市交通小区之间出租车的行程时间数据分析。从实际应用的角度出发,在交通领域对时间段的常规划分的基础上,进行交通小区之间的时间代价的研究,给出置信度时间指标。由于原数据的概率密度曲线不好拟合,所以通过建立转换模型,将原数据转换为在可接受程度内的正态分布,并使用JB统计量进行拟合效果的检测以及转换参数的选取判断。实验表明将该模型应用在行程时间数据的分析上能够很好的满足需求。3)上海市交通小区之间路径及行程时间查询平台构建。本文为了使研究的结果具有一定的完整性,在处理的数据结果集的基础之上进行了web查询平台的构建。针对数据结果之间关系以及查询展现的形式,构建数据在数据库中的存储表,同时,为了使得该平台的开发具有较好的适应性和可移植性,进行了一些接口以及软件上的优化。更多还原

【Abstract】 With the increasing popularity of the automobile,urban congestion is getting worse,more and more people pay attention to the choice of travel route.In recent years,on the one hand,the rapid development of network technology,making all aspects of data acquisition much easier,on the other hand,the rise of data mining technology has brought a boom in data research.Based on the above described,the path selection problem of urban traffic is studied by using the trajectory data,the use of taxi trajectory data in this paper.The main work of this dissertation can be summarized as the following three aspects:1)Implementation of path planning algorithm between traffic zones based on taxi trajectory data of Shanghai.Mainly aimed at the selection of taxi trajectory data mining,thus providing a new way to study the path selection.By combining with spectral clustering,a good solution to track data on the dimensions of imbalance.Analyzing and improving the current lack of spectral clustering algorithm(optimizing Gaussian kernel scale parameter optimization).Finally,the concept of local-level road network is put forward,studying complex path selection from the local road network.Experimental results show that the path classification algorithm and the path selection algorithm proposed has some enforceable.2)Travel time data analysis of taxi in Shanghai traffic zones.From the point of view of practical application,on the basis of the general division of the time period in the traffic field,research in the time cost of traffic zones and give confidence time index.Because the probability density curve of the original data is not good fitting,we establish conversion model and concert the original data to a normal distribution in the acceptable degree,what’s more,we also use the JB statistic to carry on the fitting effect the detection as well as the transformation parameter choice judgment.Experiments show that the model can be used in the analysis of travel time data to meet the demand.3)Construction of route and travel time query platform of Shanghai traffic district.In order to make the results of the study have a certain integrity,On the basis of the data result set,the web query platform is constructed.Considering the relationship between data results and the form of query display,Building data in a database storage table,at the same time,in order to make the development of the platform has a good adaptability and portability,some of the interface and software optimization are carried out.更多还原