The Design And Implementation Of Indoor Localization System Based On Wi-Fi Signal Processing

In recent years,with the prevalence of smart mobile devices,numerous smart systems and applications have emerged in our daily life,which are reshaping people's living style.With the rapid advancement in sensation,communication,and computation,indoor location based services have improved the service experience of the smart systems.Wi-Fi signal processing based fingerprint is a widely used indoor localization technology.However,the current localization frameworks based on the fingerprint localization technique face huge application challenges.First,fingerprint-based localization methods need a large amount of overhead in the offline phase,which hinders the practicability of the localization systems.Second,the current localization system based on the fingerprint cannot fully meet the application requirements for localization accuracy.Besides,Wi-Fi localization technology needs to be further improved for indoor tracking.In view of the above problems,this thesis proposes related solutions from the perspective of algorithm design and system implementation.The main contributions of this thesis are included as follows:1)we propose a framework to reduce the overhead in the offline phase for the fingerprint-based method.First,we implement a localization system prototype for the fingerprint-based method.On this basis,we propose a localization framework based on transfer learning.The basic principle is to reshape data distributions in the target domain based on the transferred knowledge from the source domains,so as to reduce the offline overhead of the target domain.Finally,extensive real-world experiments are conducted to demonstrate the effectiveness and generality of the proposed framework.2)we propose a multilateral-confined method for fingerprint-based localization to improve the localization accuracy.First,a multilateral-confined method is proposed based on Wi-Fi signal processing.Second,we implement an indoor localization system prototype based on the proposed method.Finally,comprehensive real-world experiments are conducted and the results demonstrate that the proposed solution is able to effectively improve the localization accuracy compared with conventional fingerprint-based localization techniques.3)we propose an indoor tracking system based on Wi-Fi and PDR fusion localization technique.First,we design and implement an indoor tracking system based on fusing Wi-Fi technique and PDR technique.Second,we propose a weighted fusion method based on the distance of Wi-Fi localization and PDR localization results.Finally,comprehensive real-world experiments are conducted to demonstrate the effectivity of the proposed indoor tracking system.