Indoor Positioning System Based On CSI Database

With the development of intelligent devices and wireless local area networks,people pay more attention on location-based services.Global positioning system(GPS)cannot provide precision positioning in the indoor environments.it cannot meet people's needs for indoor location services just rely on looking for schematics,asking the staff and so on.Therefore,the indoor positioning technology has become a hot research.The positioning technology used in Indoor Positioning,include ultrasound,machine vision,WIFI,radio frequency identification at home and abroad,but there are many disadvantages in positioning accuracy,equipment price,applicability and so on,there is not an indoor positioning system that meets commercial standards.People pay more attention on the WIFI-based indoor positioning technology with the deployment of indoor WIFI.The channel state information(CSI)can be obtained through the wireless network card.Compared with RSS,the channel state information(CSI)can provide more granular physical information and is more sensitive to the environment.This thesis focuses on the location fingerprint localization method based on CSI information.On this basis,this thesis proposes a weighted space-frequency characteristics clustering algorithm based on CSI.Meanwhile,this thesis designed and realized the indoor positioning system.In this thesis,the indoor location method based on WIFI position fingerprint library and the influence of different fingerprint information on the positioning effect is studied.The 802.11 n WIFI standard protocol supports multiple-input multiple-output(MIMO).The multi-antenna receives the source information at the same time and show different signal characteristics.In the OFDM transmission mode,the data is modulated at the same time on the subcarriers of different frequencies.The frequency selective fading makes the subcarriers have rich amplitude and phase information,so the subcarrier has rich amplitude and phase information.It determined that CSI data will be used as fingerprint information by simulation.Then,CSI data has the spatial and frequency characteristics,combined with CSI information and location fingerprint method,this thesis propose a clustering algorithm based on CSI weighted spatial frequency characteristics.This thesis determined the key parameters of the positioning system by simulation.CSI data is collected at the reference point,and the CSI fingerprint data is generated by data preprocessing.The clustering algorithm proposed in this thesis is used to cluster the fingerprint data and complete the matching and positioning,which greatly improves the positioning accuracy.Then,in order to further improve the positioning accuracy,this thesis proposed an indoor positioning algorithm based on the secondary correction of the fingerprint library.,the positioning accuracy is improved by removing the reference points of the database without reference value and the inappropriate selection the nearest neighbor.Finally,this thesis designed and completed the indoor positioning system based on the CSI Tools platform and wireless board.Meanwhile this designed and completed the entire indoor positioning on the second floor of the school main building.The results show that the indoor positioning system improves the positioning accuracy of the WIFI positioning technology.