Research On Indoor Localization Technology Based On WLAN

The high-precision and autonomous positioning of mobile terminals is a challenging topic and an important basis for various application service providers.With the development of mobile terminals and various location-based services,the localization of the mobile terminal is becoming more and more important.At present,GPS,Beidou and other outdoor satellite positioning systems have achieved high-precision positioning in open-air scenarios.However,because of the attenuated signal penetration and power limitations,satellite positioning systems will lead to invalid indoor localization.Therefore,indoor localization is gradually becoming the hot research in large research institutions and universities.WLAN-based indoor positioning technology has the advantages of low cost,convenient deployment,wide application,and large amount of positioning terminals.With the popularity of location-based services,its application prospects and value will be further highlighted.To improve the accuracy and robustness of WLAN-based indoor positioning technology,the main contributions of this thesis are as follows:1)In order to obtain high quality signals,the characteristics of the WiFi signal is analyzed from the representational ability of RSS,feature extraction of RSS,impact of mobile terminal orientation on received signals,etc.,so as to provide basis for obtaining high quality signal sources.2)At the stage of RSS preprocessing,this thesis presents a WiFi signal filtering algorithm based on isolated forest.This algorithm effectively filters out the abnormal value of the WiFi signal,and provides a stable and effective RSS to the following positioning algorithm.3)This thesis proposes a multi-terminal multi-stage location fingerprinting algorithm based on the improved XGBoost,which effectively improves the positioning accuracy and eliminates the influence of multi-terminal differences on the positioning accuracy.In terms of multi-terminal difference,this algorithm proposes a multi-terminal differential elimination algorithm based on gradient invariance;as for the characterization capability of WiFi signal,a wireless AP selection method based on information entropy is proposed;by analyzing the difference of local and the global representation ability of WiFi signal,an improved XGBoost multi-level fingerprint localization algorithm is elaborate.4)An indoor positioning platform based on WLAN is built.The experimental verification of the multi-terminal multi-stage location fingerprinting algorithm based on the improved XGBoost proposed in this thesis is carried out.The experimental results are compared and the effectiveness of the algorithm is verified.The results show that the average localization error decreased from 8.57m to 3.11m.