Research On Indoor Target Localization Technology Based On Modified Channel State Information

With the rapid development of the Internet of Things,emerging a large number of Indoor Location-Based Services(ILBS),which brings great convenience to daily life.Currently,indoor localization mainly relies on UWB,RFID,ultrasonic and so on.However,as indoor localization continues to be deployable,universal,and accurate,compared with Wi-Fi-based solutions,others usually require additional infrastructure.Besides,the widespread deployment of Wi-Fi infrastructure provides an opportunity for indoor localization,owning a greater cost advantage.This article proposes the following methods for the issue of Channel State Information(CSI)based Wi-Fi indoor localization:Firstly,aiming at the shortage of Received Signal Strength Indicator(RSSI)and the advantage of CSI when they are used in indoor localization,we propose an improved method for CSI error correction.Through the analysis of the process of obtaining CSI by the network card,we design different methods to eliminate CSI's errors based on the characteristics of them respectively,which lays the foundation for improving accuracy of localization.Secondly,aiming at the problem of limited resolution when the commodity Wi-Fi NIC is used to estimate the Angle of Arrival(AoA),we present a CSI-based super-resolution AoA estimation algorithm.In order to realize the virtual expansion of array elements,based on multi-subcarrier measurements,we add the phase shift introduced by the Time of Arrival(ToA)among different subcarriers to the steering vector,and then,based on the MUSIC algorithm,we implement the direct path AoA and ToA joint estimation.Thirdly,aiming at the inaccurate of the RSSI ranging for indoor environments,based on CSI filtering,we propose a parameter training-based ranging scheme.Combining CSI threshold filtering and frequency diversity,we achieve multipath suppression during ranging,and then,a parameter training based ranging model is used to map the filtered ranging parameter as distances.Combining with the estimation of AoA and ranging,we obtain the target location estimate.Finally,based on the commodity Intel 5300 Wi-Fi NIC,without any hardware changes,we extract CSI and perform the localization scheme of this paper.The experimental results in different environments show that,based on a single Access Point(AP),our scheme achieves the median localization accuracy of about 1.5m.