The Research On WIFI Fingerprint-based Localization

Recently,the widely application of Location Based Service(LBS)motivates the positioning to play a significant role in pervasive or ubiquitous computing.WIFI fingerprint-based indoor localization is the most competitive and promising positioning method taking advantage of low costs,depending on the deployed WIFI environment off the shelf,simple methodology and practical value.The paper summarized the methodology,process and mainstream algorithms of WIFI fingerprint-based indoor localization.Also,we considered the third floor of Postgraduate Building in Ningxia University as the environmental scene,using the application of WIFI Analyzer installed in the mobile device to collect the values of RSS from the surrounding AP signals,building up an experimental environment to achieve room-level WIFI localization.In the offline data collecting phase,we adopted the strongest AP method and the mean AP method respectively preprocessing the data and compared the impact on the positioning accuracy afterwards.In the online matching phase,we compared the localization performance between Bayes method which is based on possibility and NN(Nearest Neighbor)method,and we proposed an improved AP-weighted NN algorithm based on the NN algorithm,which gave different weight values to AP according to the degrees of stability to ensure that the stable AP took up a more proportion in the process of matching operation dominating the entire positioning result and diminishing the effect on positioning performance resulting from the noises.The result indicated that the AP-weighted NN algorithm was superior to the previous algorithm in terms of accuracy.Finally,considering the problem of potential location privacy leakage,we proposed a secure encryption model for the WIFI fingerprint-based indoor localization to protect the privacy of users effectively.Furthermore,we proved the accuracy and security of the safety model in theory,providing certain reference for the future research on the WIFI fingerprint-based indoor localization.