基于WiFi指纹与惯性导航的巡库员室内定位及跟踪系统的研究与实现

发布时间：2018-08-29 14:38

【摘要】：近年来,随着各种无线通信技术和网络技术的层出不穷,人们对基于位置的服务需求日益增加。全球定位系统GPS(Global Position System)的出现,满足了人们在全球范围内实现定位与导航的需求。对于信号畅通的室外环境,全球定位系统可以提供给人们精度较高的位置信息,而对于室内环境,由于建筑物本身的遮挡,使得GPS定位系统无法通过卫星来获取定位所需的信息,尤其在遮挡物较多的军火库中,更是无法利用GPS完成室内定位。因此,一种能够应用于室内环境的定位技术成为定位领域的研究重点。随着基于WLAN的WiFi定位技术和基于传感技术的惯性导航定位技术的逐步发展,它们以成本低和环境适应性强在室内定位领域得到广泛应用。同时,随着射频识别(RFID)技术的发展与兴起,其非视距、免接触、高速识别等特性也被人熟知。在实际应用环境中,大型军火库的巡库员通常配有巡库车来完成大量的巡逻任务。然而,由于监控中心无法确定巡库员的具体位置,以致部分巡库员擅离工作岗位而不被察觉,从而引发安全隐患。为了解决这一问题,本课题将设计一套巡库员室内定位系统,可以实时监控巡库员的位置并跟踪巡库员的行驶轨迹,以确保巡库员在工作时间内完成巡逻任务。本文首先通过介绍和对比目前国内外的室内定位技术以及成熟的定位系统,对不同定位技术各自的优势和局限性进行了分析。其次,针对大型军火库这一特殊的应用环境,提出了一种基于WiFi指纹与惯性导航的巡库员室内定位及跟踪系统。该系统采用C/S软件架构,由服务器端和客户机端两部分构成。首先,服务器端完成巡库员定位信息的采集,服务器端的核心是STM32F103微控制器,它连接WiFi、地磁传感器、霍尔传感器、RFID四大模块,分别采集巡库员的WiFi指纹、行驶方向、行驶距离以及经过的射频标签;紧接着,服务器端将所有定位信息通过基于TCP/IP的Socket通信发送到客户机端;最后,客户机端在离线阶段采用均值平滑法来减小指纹序列的波动性,并采用K-means聚类算法对WiFi指纹进行分类处理;在在线定位阶段,通过改进后的K近邻指纹算法与惯性导航的融合完成定位,并在搭建的系统测试平台上实时定位巡库员的位置并对其行驶轨迹进行跟踪。
[Abstract]:In recent years, with the emergence of various wireless communication and network technologies, the demand for location-based services is increasing. Global Positioning system (GPS (Global Position System) meets the needs of global positioning and navigation. For outdoor environments with unblocked signals, GPS can provide people with more accurate location information, while for indoor environments, due to the occlusion of the building itself, The GPS positioning system is unable to obtain the information needed by satellite, especially in the military depot where there are more occluded objects, so it is impossible to use GPS to complete the indoor positioning. Therefore, a kind of localization technology which can be applied to indoor environment has become the focus of research in the field of localization. With the development of WiFi positioning technology based on WLAN and inertial navigation positioning technology based on sensor technology, they are widely used in indoor positioning field because of their low cost and strong environmental adaptability. At the same time, with the development and development of RFID (RFID) technology, its non-line-of-sight, contactless, high-speed identification and other characteristics are also well known. In practical applications, large armoury inspectors are usually equipped with patrol vehicles to complete a large number of patrol tasks. However, because the monitoring center can not determine the specific location of the inspectors, some of the inspectors leave their jobs undetected, thus causing security risks. In order to solve this problem, a set of indoor positioning system is designed, which can monitor the position of the inspector in real time and track the track of the patrol, so as to ensure the patrol task is completed during the working hours. In this paper, the advantages and limitations of different positioning techniques are analyzed by introducing and comparing domestic and foreign indoor positioning technology and mature positioning system. Secondly, aiming at the special application environment of large armoury, an indoor positioning and tracking system based on WiFi fingerprint and inertial navigation is proposed. The system uses C / S software architecture and consists of two parts: server and client. First of all, the server completes the acquisition of the information of the inspector's location. The core of the server is the STM32F103 microcontroller, which connects the WiFi, geomagnetic sensor and the Hall sensor to the four modules of WiFi, and collects the WiFi fingerprint of the inspector and the driving direction, respectively. Then, the server sends all the location information to the client through the Socket communication based on TCP/IP. Finally, the client uses the mean smoothing method to reduce the volatility of the fingerprint sequence in the off-line phase. K-means clustering algorithm is used to classify the WiFi fingerprint, and in the online location stage, the improved K-nearest neighbor fingerprint algorithm and inertial navigation are combined to complete the localization. And the real-time location of the inspector is located on the system test platform and the track of the patrol is tracked.
【学位授予单位】：东华大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN92;TN96

【相似文献】