基于高斯过程回归的Wi-Fi RTT/RSS测距与指纹定位研究  

Wi-Fi RTT/RSS ranging and fingerprint positioning research based on Gaussian process regression

作　　者：谢思语 王鑫龙 邱燕华 李彤云 师嘉怡 汪云甲[1] 陈国良[1] 孙猛 XIE Siyu;WANG Xinlong;QIU Yanhua;LI Tongyun;SHI Jiayi;WANG Yunjia;CHEN Guoliang;SUN Meng(School of Environment and Spatial Informatics,China University of Mining and Technology,Xuzhou,Jiangsu 221116,China)

机构地区：[1]中国矿业大学环境与测绘学院,江苏徐州221116

出　　处：《导航定位与授时》2024年第5期36-52,共17页Navigation Positioning and Timing

基　　金：国家自然科学基金(42304047,42274048);江苏省重点研发计划项目(BE2022716);中国矿业大学省级大学生创新创业训练计划项目(202310290176Y)。

摘　　要：基于往返时间(RTT)测量的智能手机Wi-Fi测距定位受限于室内环境的复杂性,仍面临稳定性差、精度低等问题。利用同步量测的Wi-Fi RTT和信号接收强度(RSS)数据,分别从测距与指纹补偿、测距定位与匹配定位优化等方面开展研究。首先,通过分析RTT测距误差规律,建立了基于高斯过程回归(GPR)的非参数测距误差补偿模型;研究了RSS数据分布,通过拟合Wi-Fi信号路径衰减模型,构建了基于GPR的RSS补偿模型。其次,开发了基于Web端的指纹库生成和指纹定位软件,可支持RSS指纹库、RTT测距指纹库自主建设和RSS/RTT指纹定位。最后,设计了基于GPR补偿的RTT测距定位、RTT指纹定位和Wi-Fi RSS指纹匹配定位算法,并综合分析了3种方法的定位性能。实验结果表明,经过高斯补偿的RTT测距定位、RTT指纹定位和RSS指纹定位的平均精度分别提升了50.81%、52.85%和48.72%,证明了高斯过程回归模型可有效提升Wi-Fi RTT/RSS测距与指纹定位的精度与稳定性。Smartphone-based Wi-Fi ranging positioning using round-trip time(RTT)measurement is limited by the complexity of the indoor environment and still faces problems such as poor stability and low accuracy.Simultaneously measured Wi-Fi RTT and received signal strength(RSS)data are used to conduct research from the aspects of ranging and fingerprint compensation,ranging positioning and fingerprint optimization.Firstly,a non-parametric ranging error compensation model using Gaussian process regression(GPR)is established by analyzing the ranging errors.The path loss model of Wi-Fi signal propagation is fitted based on the RSS distribution research,and by fitting the Wi-Fi path loss model,an RSS compensation model based on GPR is constructed.Second,a Web-based fingerprint database generation and fingerprinting software is developed,which can support the independent construction of RSS fingerprint database and RTT ranging fingerprint databases and RSS/RTT fingerprint positioning.Finally,RTT ranging positioning,RTT fingerprinting,and Wi-Fi RSS fingerprint-matching positioning algorithms based on GPR compensation are designed,and the positioning performance of the three methods is comprehensively analyzed.The experimental results show that the average accuracy of RTT ranging positioning,RTT fingerprint positioning and RSS fingerprinting with GPR compensation has increased by 50.81%,52.85%and 48.72%,respectively,proving that the GPR model can effectively improve the positioning accuracy and stability of Wi-Fi RTT/RSS ranging and fingerprint positioning.

关 键 词：室内定位 高斯过程回归 Wi-Fi精细时间测量 往返时间 指纹定位 测距定位 

分 类 号：P227[天文地球—大地测量学与测量工程]