基于智能手机GPS的大学生出行方式识别研究

发布时间：2018-01-08 17:35

本文关键词：基于智能手机GPS的大学生出行方式识别研究　出处：《江苏大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着我国高等教育规模的快速发展,高等院校迁至城郊结合部甚至远郊已成为普遍现象,但城郊结合部的交通基础设施很难满足高校师生的出行需求。为降低大学生出行对校园周边交通乃至城市交通网络的影响,需对交通网络进行优化设计,对大学生进行出行调查是必须开展的一项基础性工作。传统出行调查方法主要依赖被访问者对行程的回忆及其主观认知,调查数据往往存在较多缺陷:如被访问者的负担较重、访问回应率低、数据质量差、存在漏报错报等。基于便携式GPS的出行调查同样存在一定缺陷:调查成本高和因忘带便携式GPS而导致出行数据遗漏等。随着手机技术的不断发展,GPS定位系统已成为智能手机标配。如今大学生出门携带智能手机已成为日常习惯。因此,基于智能手机GPS的大学生出行调查不仅能够降低出行调查费用,而且能减少出行数据遗漏现象的发生。从智能手机GPS记录的出行轨迹数据中提取出行信息以及识别出行方式成为分析大学生出行行为一种新途径。(1)本文先对大学生出行方式调查方案及实施方案进行研究,认真分析其优缺点,为大学生出行轨迹数据处理与出行方式识别提供数据保障。(2)对利用智能手机GPS收集的大学生出行轨迹数据进行处理:首先是对轨迹数据进行预处理,包括数据过滤和数据格式转换;然后是在GPS信号缺失情况下选择停留时间和平均速度这两个参数,在GPS信号正常情况下选择临界距离、最小停留时间和最大停留时间这三个参数,基于混合方法利用这五个参数进行出行段识别;最后提取出行特征变量,并利用箱线图法和组间均值等式检验法验证其有效性。(3)对大学生出行方式进行识别。针对粒子群算法容易早熟收敛的缺陷,本文利用改进粒子群来优化支持向量机。然后利用IPSO-SVM模型对步行、自行车、电动车、校园公交、公交车和出租车进行识别研究。选取线性核函数、多项式核函数和径向基核函数分别作为SVM的核函数,得出不同核函数下IPSO-SVM模型的出行方式识别精度,并选择平均识别精度最高的作为IPSO-SVM模型最终识别精度。将IPSO-SVM模型识别精度与其他常用出行方式识别模型的识别精度进行对比。研究结果表明,本文提出的IPSO-SVM模型在基于智能手机GPS的大学生出行方式识别研究中具有更好的识别精度。研究结论对智能手机GPS在大学生出行研究领域的推广,科学地分析大学生的出行规律,诊断高校周边的交通问题,促进高校周边交通系统的健康发展具有深远意义。
[Abstract]:With the rapid development of China's higher education scale, it has become a common phenomenon that institutions of higher learning move to suburban areas or even suburbs. In order to reduce the impact of college students' travel on campus traffic and even urban traffic network, the traffic network should be optimized. It is necessary to carry out a trip survey for college students. The traditional travel survey method mainly depends on the visitors' recollection of the trip and their subjective cognition. Survey data often have more shortcomings: such as the heavy burden of the interviewee, low response rate and poor data quality. There are false reports and so on. The trip survey based on portable GPS also has some defects: high cost of investigation and omission of travel data due to forgetting portable GPS. With the development of mobile phone technology. The GPS location system has become the standard for smartphones. Nowadays, it is a daily habit for college students to go out and carry their smartphones. The trip survey of college students based on smart phone GPS can not only reduce the cost of travel survey. It can also reduce the occurrence of trip data omission. Extracting trip information from trip track data recorded by GPS of smart phone and identifying trip mode become a new way to analyze college students' trip behavior. In this paper, first of all, the investigation scheme and implementation plan of college students' travel mode are studied. Carefully analyze its advantages and disadvantages. Provides data guarantee for the data processing and identification of college students' trip path. (2) processing the data of college students' trip path collected by smart phone GPS: firstly, preprocessing the path data. Including data filtering and data format conversion; Then the two parameters of residence time and average velocity are selected in the absence of GPS signal, and the three parameters of critical distance, minimum residence time and maximum residence time are selected under the normal condition of GPS signal. The five parameters are used to identify the travel segment based on the hybrid method. Finally, the travel characteristic variables are extracted, and the validity of the method is verified by the box-line graph method and the mean equality test method between groups. The paper aims at the shortcomings of particle swarm optimization (PSO), which is easy to converge prematurely. This paper uses improved particle swarm optimization (PSO) to optimize support vector machine (SVM). Then the IPSO-SVM model is used for walking, bicycle, electric vehicle and campus bus. The linear kernel function, polynomial kernel function and radial basis kernel function are selected as the kernel functions of SVM. The identification accuracy of IPSO-SVM model under different kernel functions is obtained. The recognition accuracy of IPSO-SVM model is compared with that of other common travel modes. The results show that. The IPSO-SVM model proposed in this paper has better recognition accuracy in the research of college students' trip pattern identification based on smart phone GPS. The conclusion of the research is that the smart phone GPS can be used in the field of college students' travel research. Promotion. It is of great significance to scientifically analyze the travel rules of college students, to diagnose the traffic problems around colleges and universities, and to promote the healthy development of traffic system around colleges and universities.
【学位授予单位】：江苏大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：U491

【参考文献】