Research On Indoor Localization Technology Based On IBeacon And Fingerprint

In recent years,with the development of Bluetooth 4.0,especially the maturity of iBeacon technology proposed by Apple in the near field communication,the high precision indoor localization based on the Bluetooth Low Energy 4.0 technology has gradually become a new hotspot in the area of indoor localization.At present,indoor localization system based on iBeacon technology is yet being widely used,and most of the system is still on the theoretical stage and under testing.The current indoor localization system is still rare,lacking of mature of commercial applications.Under the existing technology and theory of fingerprint localization,this thesis tries to put forward one kind of solution based on iBeacon technologyto meet engineering performance requirements.The main description of this thesis is as follows:First,in order to solve the problem caused by the RSSI jumpingwhich due to the environment interference,we put forward the strategy of Gauss filtering and Kalman filtering to measure the RSSI of the measurement,eliminate the error caused by random interference,and ensure the accuracy of the RSSI data acquisition source.Second,in order to reduce the influence on the localization performance caused by RSSI,it is necessary to study the distribution characteristics of RSSI at the reference point.This thesis analysis the influence on RSSI distribution caused by the several factors which are distance,personnel occlusion and different personnel density by means of mathematical statistics and histogram from a quantitative point of view,and study the relationship of RSSI from different Beacons.We concluded that the RSSI satisfies the Gauss distribution characteristics in a room.Third,based on the above analysis,this thesis aims to improve the performance of the location system,begins with the research on the fingerprint localization from two aspects of efficiency and precision.In terms of efficiency,this paper proposes a preprocessing method of K-means clustering algorithm,which deals with the fingerprint data from offline database and the testing point to solve the huge,time –consuming issue.Compared with the non-clustering system,localization time decreased significantly.In terms of precision,according to the limitation of the traditional naive Bayesian probabilistic algorithm that does not consider the RSSI irregular distribution,this thesis proposes a Gaussian Mixture Model to compute the probability,the Mixing model parameters obtained by the EM algorithm,and we can reduce the error of positioning results and improve the positioning accuracy.Fourth,combining with the sample data which are collected in the off-line stage,we can construct the off-line fingerprint database.The simulation experiments on the MATLAB platform proved thatthe method of using K-means clustering and improved probabilistic algorithm for localization is feasible and effective,providing a reference for the application of iBeacon technology in the actual development.