基于无线信道状态相位信息优化的定位算法

随着位置服务需求的增长,基于Wi-Fi接收信号的室内定位技术一直是研究热点之一.通过检测环境变化对Wi-Fi无线信道状态信息CSI的影响,从而实现对室内人员的定位具有通用性强、部署成本低等优点.针对大多系统仅使用CSI中幅度信息所带来准确性和稳定性不足的问题,设计并实现了一种基于CSI相位信息优化的定位算法,该方法通过采集幅度和相位参数相结合作为位置指纹特征,并对特征数据进行预先平滑去噪后进行指纹库的构建,然后通过机器学习方法进行人员位置的分类识别.由于相位和幅度信息可以相互补充,弥补了某些易混淆位置的分类错误,从而解决了采用单一特征的定位准确性和稳定性问题.实验进行了两种不同多径场景下的实验,比较了不同指纹特征选取、数据预处理方法以及三种机器学习算法对定位准确度的影响,其结果表明采用本文所提出算法总体上可以在仅使用CSI幅度特征的基础上提高13%.

Localization Algorithm based on Wireless Channel State Phase Information Optimization

With the increasing demand for location services,indoor localization technology based on Wi-Fi has been one of the research hotspots. By detecting the impact of environmental changes on the channel state information ( CSI) ,Wi-Fi generally has the advantages such as strong versatility and low deployment cost. For the problem that most systems only use the amplitude feature of CSI,this paper designs and implements a localization algorithm based on the CSI phase feature optimization. This method combines the amplitude and phase parameters as a position fin-gerprint. Fingerprint databases were constructed after pre-smoothing and denoising the feature data,and then the classification of personnel positions was identified by machine learning methods. Because the phase and amplitude information can complement each other,it can make up for the misclassification of certain confusing positions,thus solving the problem of positioning accuracy and stability using a single feature. Experiments were performed in two different multipath scenarios. The effects of different fingerprint feature selections,data preprocessing methods,and three machine learning algorithms on positioning accuracy were compared. The results show that the proposed algo-rithm can generally increase by 13% on the basis of using only CSI amplitude characteristics.