基于位置指纹的室内无线定位技术研究

发布时间：2018-07-05 19:00

本文选题：定位技术 + 无线局域网　；参考：《兰州交通大学》2016年硕士论文

【摘要】：随着互联网的高速发展和移动智能终端的普及应用,人们对基于位置服务(Location Based Services,LBS)的需求日益增加。目前LBS已经被广泛应用于社会生活和工业生产的多个领域,LBS应用中最关键的部分是对位置信息的获取,即定位技术,定位技术的好坏直接影响着LBS的应用效果。尽管卫星导航定位技术比较成熟,能够满足大多数室外位置服务的需求,但当将其应用在室内时并不能取得理想的效果,因此对室内定位技术的研究越来越受到人们的关注。现有的室内定位技术大多数都需要相关的专用硬件设备,使得定位成本高、应用不灵活,很难得到大范围的应用和推广。基于无线局域网(Wireless Local Area Networks,WLAN)的室内定位技术无需专用的硬件设备,仅利用相应的软件便可通过移动智能终端进行定位,定位成本低并能满足大多数室内定位精度的要求,因此WLAN定位技术成为了室内定位的首选。在WLAN室内定位中,位置指纹定位算法以定位精确度高、抗干扰能力强、定位成本低、支持多终端设备等优点成为应用最广泛的定位算法。尽管如此,位置指纹定位仍存在尚待完善的部分,如RSS时变性会造成定位精度下降和指纹库过期等。本文针对该问题给出了相应的解决方案,通过实验对比证明了方案的可行性。本文主要工作如下:(1)为能真实的反映RSS分布情况,提出了一种基于混合高斯分布模型的位置指纹定位算法。针对不同的RSS概率分布,采用相对应的分布模型对其进行曲线拟合,拟合曲线更符合真实的RSS分布,因此构建的位置指纹数据库更加可靠。借鉴重叠度在地图匹配和肤色检测中的应用,使用两个指纹特征RSS分布的重叠度表示两个位置指纹之间的相似度,对与待测点指纹相似度最大的前K个参考点,依据相似度大小再次分配不同权值进行质心加权运算,以估计待测点位置。(2)针对RSS的时变特性导致位置指纹库过期的问题,本文提出一种利用用户反馈信息对指纹库进行更新和维护的方法。用户反馈的信息包括自身所在位置和该点采集的各AP信号强度,依据打分机制对定位点指纹所包含的AP进行打分,删除关闭的AP和添加开启的AP以对指纹库进行更新。考虑到定位结果与用户实际位置不一致的情况,允许用户更正自身位置,并用聚类检测方法判断用户更正结果是否可信。实验结果表明,利用用户反馈的信息更新指纹库比不更新和人工低频更新指纹库的定位效果更好。
[Abstract]:With the rapid development of the Internet and the widespread application of mobile intelligent terminals, the demand for location based Services (LBS) is increasing day by day. At present, LBS has been widely used in many fields of social life and industrial production. The most important part of LBS application is the acquisition of location information, that is, location technology, which directly affects the application effect of LBS. Although satellite navigation and positioning technology is mature and can meet the needs of most outdoor location services, it can not achieve ideal results when it is applied indoors, so more and more attention has been paid to the research of indoor positioning technology. Most of the existing indoor positioning technology needs special hardware equipment, which makes the positioning cost high, the application is inflexible, and it is difficult to be widely used and popularized. The indoor positioning technology based on Wireless Local Area Network (WLAN) does not need special hardware equipment. It can be located by mobile intelligent terminal only by using the corresponding software. The positioning cost is low and can meet the requirements of most indoor positioning accuracy. Therefore, WLAN positioning technology has become the first choice of indoor positioning. In WLAN indoor positioning, location fingerprint location algorithm has become the most widely used location algorithm because of its high accuracy, strong anti-interference ability, low positioning cost and support for multi-terminal equipment. In spite of this, there are still some parts to be improved in location fingerprint location, for example, RSS time-varying will result in the decrease of location accuracy and the expiration of fingerprint database. In this paper, the corresponding solution to this problem is given, and the feasibility of the scheme is proved by experiment. The main work of this paper is as follows: (1) in order to reflect the distribution of Gao Si, a location fingerprint location algorithm based on hybrid Gao Si distribution model is proposed. According to the different RSS probability distribution, the corresponding distribution model is used to fit the curve, the fitting curve is more consistent with the real RSS distribution, so the location fingerprint database is more reliable. Based on the application of overlap degree in map matching and skin color detection, the overlap degree of RSS distribution of two fingerprint features is used to represent the similarity between two position fingerprints. According to the similarity magnitude, different weights are assigned to weight to estimate the location of the point to be measured. (2) due to the time-varying nature of RSS, the fingerprint database of position is out of date. In this paper, a method of updating and maintaining fingerprint database using user feedback information is presented. The feedback information includes the position of the user and the intensity of each AP signal collected at this point. According to the scoring mechanism, the AP contained in the fingerprint of the location point is graded, the closed AP is deleted and the open AP is added to update the fingerprint database. Considering that the location result is not consistent with the user's actual location, the user is allowed to correct his own position, and the clustering detection method is used to judge whether the user's correction result is credible or not. The experimental results show that the location effect of the fingerprint database updated by the user feedback is better than that of the unupdated fingerprint database and the artificial low-frequency fingerprint database.
【学位授予单位】：兰州交通大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TN925.93

【参考文献】