高精度室内定位方法研究与实现

发布时间：2018-05-17 14:41

本文选题：室内定位 + 信道状态信息　；参考：《北方工业大学》2017年硕士论文

【摘要】：室内的精确和实时的定位具有巨大的市场价值和重要的应用前景。由于GPS在室内环境受到诸多因素影响无法提供高精度定位,如何精确定位室内环境下的位置成为了研究和应用的热点。随着无线局域网的广泛部署,利用现有的无线网络作为基础设备,使得室内定位能够实现无需特定额外设备的、较低成本的精确定位。近年来,有研究人员采用了比RSSI更细粒度的物理层特征CSI实现定位。本文提出了一种基于稀疏表示的CSI的室内定位方法。系统包括了离线和在线两个阶段,离线用于采集训练CSI数据,而在线阶段用于实时定位。为了提高在线定位速率,降低算法的计算复杂度,离线阶段在采集完CSI后,增加了CSI指纹训练过程,转移一部分计算量到离线训练过程中。定位方案结合指纹匹配方法,弥补了以往的室内定位方案精度不高的不足。利用稀疏表示算法,减弱了CSI指纹中噪声和不稳定变量对定位精度的影响,提高了定位精度。为了验证CSI和稀疏表示算法更适用于室内定位,使用普通路由作为信号发射器,同时定制电脑系统和无线网卡驱动搭建实验环境。最后通过使用MATLAB对本文方案进行仿真测验。测验结果表明,本文算法能够有效实现无需用户携带其他特定设备的室内定位,平均精度能够达到0.12m,准确度能够达到91%。
[Abstract]:Indoor accurate and real-time positioning has great market value and important application prospect. Due to the influence of many factors on the indoor environment, GPS can not provide high precision positioning, so how to accurately locate the position in indoor environment has become a hot spot in research and application. With the wide deployment of WLAN, using the existing wireless network as the basic equipment, indoor positioning can be achieved without specific additional equipment, low-cost accurate positioning. In recent years, some researchers have adopted a finer grained physical layer feature CSI than RSSI to achieve localization. In this paper, an indoor location method based on sparse representation CSI is proposed. The system consists of two stages, offline and online, which are used to collect training CSI data, while on-line stage is used to locate in real time. In order to improve the speed of online location and reduce the computational complexity of the algorithm, the CSI fingerprint training process is added after the CSI is collected in the off-line phase, and a part of the computation is transferred to the off-line training process. The localization scheme combined with fingerprint matching makes up for the lack of accuracy of previous indoor localization schemes. By using sparse representation algorithm, the influence of noise and unstable variables in CSI fingerprint on location accuracy is reduced, and the location accuracy is improved. In order to verify that CSI and sparse representation algorithms are more suitable for indoor positioning, common routing is used as signal transmitter, and computer system and wireless network card driver are customized to build experimental environment. Finally, we use MATLAB to test the scheme of this paper. The test results show that the proposed algorithm can effectively realize indoor positioning without any other special equipment carried by the user. The average accuracy can reach 0.12 mand the accuracy can reach 91m.
【学位授予单位】：北方工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN92

【参考文献】