Research On WiFi Location Technology For Complex Scenes

With the development of wireless technology and the popularization of smart devices,location-based services show an explosive growth trend.It has a wide range of applications in navigation,rescue,tracking and other fields.WiFi-based fingerprint localization has been widely studied and applied because of its low cost,no additional hardware,wide applicability and high localization accuracy.However,WiFi fingerprint signals have volatility and are easily affected by the environments,so the localization performance will be significantly reduced in complex environments.In view of the above problems,this thesis analyzes the characteristics of WiFi fingerprint signals,and studies the WiFi localization technology in complex environments.From the two aspects of locating the internal structure of the fingerprint and the adaptiveness of the locating fingerprint environments,a spatial matrix fingerprint algorithm based on the stability of neighbors and an adaptive fingerprint location algorithm based on signal distribution are respectively proposed.The main contents are as follows:1.This thesis analyzes WiFi signal characteristics by collecting fingerprint data,and summarizes feature that affects the localization effect in complex scenarios.The Received Signal Strength(RSS)of the WiFi signal is used to analyze the volatility and ambiguity of the RSS signal.Then,the stability change patterns of the RSS signal is explored through the analysis of the signal characteristics of the reference point and its neighboring points.Further analyze the regional differences of the RSS signal in different stable scenes,find that the effectiveness of fingerprint localization of different structures is different and analyze its impact.2.In order to solve the ambiguity problem caused by the fluctuation of RSS mean fingerprint,considering the stability of fingerprint spatial nearest neighbor,a spatial matrix fingerprint algorithm based on the stability of nearest neighbor is proposed to reduce the proportion of ambiguous points and improve the localization accuracy.Firstly,the fingerprint information of the reference point and its surrounding neighbor points are constructed as a neighbor matrix fingerprint based on the characteristics of neighbor stability.The data relationship between the neighbor fingerprints is used to repair the missing value of the data lost during the collection process,and the stability of the matrix fingerprint is performed quantification.The similarity and stability of the joint matrix fingerprints are used as the standard,and two similarity measures are used to characterize the similarity,and the appropriate reference points are screened out.Finally,the K-nearest neighbors are calculated to match the reference points to complete the coordinates of the final localization of the target estimate and perform performance analysis.3.This thesis also aims at the problem of the different applicability of RSS fingerprints of different structures to different localization scenarios under different stationary states in complex environments.An adaptive fingerprint localization algorithm based on signal distribution is proposed,which considers the adaptability of localization fingerprints.First of all,through inspection and analysis,the stability characteristics of the fingerprint itself and the stability characteristics of the relationship between the neighbors are perceived,and the relationship between the stability in the localization scene and the selection of the localization fingerprint is found.The self-stationarity coefficient σ and the nearest neighbor stability parameter β are introduced to adjust.The number of selected reference points for different fingerprint types determines the final localization reference point ratio.Finally,the K-nearest neighbor algorithm is used to estimate the target position and perform performance analysis.