基于众包的嵌套流形匹配室内定位方法

发布时间：2018-03-18 18:27

本文选题：室内定位　切入点：众包　出处：《新疆大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着普适应用的兴起,室内定位也变得越来越重要。在大场景或者陌生场景中,用户最关心的是自己的位置。受益于无线网络的快速发展、无线接入技术的流行,基于WiFi进行室内定位成为主流。但传统的指纹定位需现场勘测采集数据,所需时间及工作量巨大,且当室内变化时所建指纹库会“失效”。在保证指纹库时效性和准确性的同时,降低人工及时间成本,使其成为一种普适性的室内定位技术,成为该领域全新的研究方向,也是该领域研究者的一个挑战。本文采用众包的形式进行室内位置信息的采集,并记录用户的大量路径信息,利用嵌套在路径中的低维流形一致性进行位置匹配以建立指纹库。通过高斯粒子滤波对传感器数据去噪,解决步长差异问题。定位时,根据用户位置连续性和路径信息筛选合理近邻点,继而实现精确定位。实验表明在不需要现场勘测的情况下,可达到和传统方法可比的定位精度。该方法可实时适应环境变化,经过2周甚至1个月之后,定位准确性更是优于传统定位方法。本文主要解决了以下四个方面的问题:·在保证定位误差不是很大的情况下,通过众包的形式大大减少位置指纹库建立所消耗的时间以及人力等;·解决了传统数据库不能适应室内变化的问题,可以根据新采集的众包数据更新指纹库,以此来保证其可适应性;·采用高斯粒子滤波器对传感器数据进行去噪,进而计算已知步数情况下行走距离,求得各行人步长后归一化处理,以解决步长差异问题;·在线定位时,利用位置的连续性和库内详细路径信息进行近邻点选择,有效排除“伪近邻点”,从而降低定位误差。
[Abstract]:With the rise of pervasive applications, indoor positioning is becoming more and more important. In large or unfamiliar scenarios, users are most concerned about their own location, benefiting from the rapid development of wireless networks and the popularity of wireless access technology. Indoor location based on WiFi becomes the mainstream. But the traditional fingerprint location needs to survey and collect data on the spot, the time and workload are huge, and the fingerprint bank will be "invalid" when the room changes. At the same time, it can ensure the timeliness and accuracy of the fingerprint database. Reducing the cost of labor and time, making it a universal indoor positioning technology, has become a new research direction in this field, and is also a challenge for researchers in this field. In this paper, crowdsourcing is used to collect indoor location information. A lot of path information of the user is recorded and position matching is carried out by using the consistency of the low dimensional manifold nested in the path to establish the fingerprint database. The sensor data is de-noised by Gao Si particle filter to solve the problem of the difference in step size. According to the user position continuity and path information, the reasonable nearest neighbor points are screened, and the accurate location is realized. The experiment shows that the location accuracy can be comparable to that of the traditional method without the need of site survey. This method can adapt to the environment change in real time. After two weeks or even a month, the positioning accuracy is better than the traditional positioning method. This paper mainly solves the following four problems: 路ensure that the positioning error is not very large, By crowdsourcing, the time and manpower consumed in the establishment of the location fingerprint database are greatly reduced; 路the problem that the traditional database cannot adapt to indoor changes can be solved, and the fingerprint database can be updated according to the newly collected crowdsourcing data. In order to ensure its adaptability; 路using Gao Si particle filter to Denoise the sensor data, and then calculate the walking distance under the known number of steps, and then get the step size of each pedestrian, then normalize the step size to solve the problem of step size difference; 路online positioning, The location continuity and the detailed path information in the database are used to select the nearest neighbor points, and the pseudo nearest neighbor points are effectively excluded, thus reducing the location error.
【学位授予单位】：新疆大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN92

【相似文献】