基于WLAN位置指纹的室内定位技术研究

发布时间：2018-04-06 07:12

本文选题：室内定位　切入点：位置指纹　出处：《南京邮电大学》2017年硕士论文

【摘要】：随着无线通讯技术的高速发展和各式各样智能终端的大范围使用,用户对拥有高准确性的室内定位服务的需求在急剧增长,在治病就医、抗震救灾、公共社交媒体、交通工具导航和场景监控等领域都展现出了巨大的市场远景。基于WLAN位置指纹室内定位算法实现起来相对简单,构建起定位环境的成本较低,符合在绝大多数室内定位中应用的需求,成为当下室内定位研究中的热门技术。本文对基于WLAN的位置指纹定位技术进行了研究。首先论述了基于WLAN的位置指纹定位算法中普遍存在的误差及其产生原因,并总结了已有的相应解决方案。本文从室内定位中离线采样和在线定位两个阶段着手,从AP定位性能差异性分析、指纹数据库的构建、数据库聚类分块和匹配定位四个主要方面进行了深入分析和研究。以降低建立位置指纹数据库时的工作量和同时提高定位精度为目标,提出了一种基于AP信号方差的位置指纹室内定位的改进算法。由于AP的分布位置不同及AP自身的差异性,使得在定位过程中不同AP对定位效果的影响不同,该方法区分了不同AP在定位过程中的贡献大小,仿真实验表明,基于AP信号方差的改进定位算法不但缩短了指纹数据库的建立时间还提高了定位的精度。其次,重点分析了K均值聚类方法在基于WLAN的位置指纹室内定位系统中的应用,总结分析了该聚类算法在定位过程中的利弊,并从理论的角度提出了相应的改进方案。最后,因为K均值聚类方法中存在类与类边界处定位性能差的问题,不能对类与类边界处参考点进行准确的选取,这直接导致在类与类相邻处的定位精度较低。为解决这个问题,本文提出了基于二次K均值聚类的位置指纹室内定位方法,通过对第一次聚类结果的再处理,减少了因类与类相邻处参考点选择不合理而带来的误差,提高了定位的精度。
[Abstract]:With the rapid development of wireless communication technology and the wide use of all kinds of intelligent terminals, the demand of users for indoor positioning services with high accuracy is increasing dramatically, and they are seeking medical treatment, earthquake relief, public social media,Vehicle navigation and scene monitoring and other areas have shown a huge market perspective.Based on WLAN location fingerprint indoor location algorithm is relatively simple to achieve, the cost of building a location environment is relatively low, in line with the needs of most indoor positioning applications, it has become a hot technology in the research of indoor location.In this paper, the location fingerprint location technology based on WLAN is studied.In this paper, the errors and their causes in the location fingerprint location algorithm based on WLAN are discussed, and the corresponding solutions are summarized.This paper starts with the two stages of off-line sampling and on-line positioning in indoor positioning, and makes in-depth analysis and research on four main aspects of AP location performance difference analysis, fingerprint database construction, database clustering and matching location.In order to reduce the workload of establishing location fingerprint database and improve the location accuracy, an improved location fingerprint indoor location algorithm based on AP signal variance is proposed.Because of the different distribution of AP and the difference of AP itself, the effect of different AP on the localization effect is different. The method distinguishes the contribution of different AP in the positioning process.The improved location algorithm based on AP signal variance not only shortens the time of establishing fingerprint database, but also improves the accuracy of location.Secondly, the application of K-means clustering method in the location fingerprint indoor location system based on WLAN is analyzed, the advantages and disadvantages of the clustering algorithm in the localization process are summarized and analyzed, and the corresponding improvement scheme is put forward from the theoretical point of view.Finally, because the K-means clustering method has the problem of poor localization performance at the boundary of classes and classes, it is impossible to select the reference points at the boundary of classes and classes accurately, which directly leads to the low positioning accuracy at the adjacent areas of classes and classes.In order to solve this problem, a location fingerprint indoor location method based on quadratic K-means clustering is proposed in this paper. By reprocessing the results of the first clustering, the error caused by the unreasonable selection of reference points between the cluster and the cluster is reduced.The accuracy of positioning is improved.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN925.93;TP311.13

【参考文献】