Smartphone-based Wi-Fi ranging positioning using round-trip time(RTT)measurement is lim-ited 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 con-duct 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 in-dependent 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 experi-mental results show that the average accuracy of RTT ranging positioning,RTT fingerprint posi-tioning 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
Key words
Indoor positioning/Gaussian process regression/Wi-Fi fine timing measurement(FTM)/Round-trip time(RTT)/Fingerprint positioning/Ranging positioning
维普
万方数据