Research Of WIFI Indoor Location Technology Based On Position Fingerprint Method

By the quickly innovation of science and technology,people can not stand away from the requirement of location information in work and daily life,the requirement of indoor location is increasing.Nowadays WIFI indoor location technology is a classical and proven indoor location technology,there are some characteristics of WIFI location,it is convenient for location and system of WIFI location is very easy,do not need some extra equipment,so that WIFI location is deeply researched in this paper.In this paper,by research of WIFI indoor location based on position fingerprint method,author finds that there are several causes affect the precision of traditional WIFI indoor location,influence factors are as follows,first factor is fluctuation of collecting signal,this results from environment influence and result in low accuracy of location datum.Second factor is that there needs a large amount of traverses in fingerprint database and make algorithm complex by the location process.The last factor is that the error range of traditional location algorithm result is much too bigger compares with actual location result.All these factors are the problems which need to be solved by this paper.Firstly author analyses precondition filtering algorithms and applies them to signal strength processing,comparing different influence of location precision before filtering and after filtering.After establishing fingerprint database,come out with a algorithm based on neighborhood average filtering-neighborhood filtering algorithm based on AP signal strength and apply this algorithm to process database in order to filter noise in fingerprint database;After precondition,considering that there exists discrete points in fingerprint database by process of classic k-means clustering algorithm,coming out with an improvement of clustering algorithm-binary k-means based on dual similarity and applying it in database processing to eliminate noise point;Finally propose improvements of location algorithm,one is AP weighted similarity WKNN algorithm,another is improved adaptive WKNN algorithm.They solve the problem that WKNN do not pay different attention to different AP and relief the influence by selecting quantity of reference point.