Design And Implementation Of WiFi Location Algorithm Based On Evidence K-Nearest Neighbor And Grid Clustering

With the rapid development of mobile network technology,indoor positioning technology has attracted more and more attention.Due to the wide application of WiFi technology,it has become a hotspot in the field of indoor positioning technology research.However,there bring difficulties to the WiFi positioning technology ascribed to complex indoor environment and numerous disturbance factors.In order to reduce the influence of all kinds of interference on WiFi localization,Localization method based on the fingerprint localization algorithm is ideal at present.However,the traditional fingerprint method based on the receiving signal strength indicator(RSSI)has the problem of large amount of fingerprint collection,low positioning efficiency and positioning accuracy.With the development of indoor positioning technology,users have different positioning requirements in different regions.How to design a widely applicable,efficient and high-precision positioning algorithm to meet the different require--ments is a problem worth studying.For the problems of the traditional the fingerprint localization method is inefficient and user demand for indoor local area positioning is different,this paper proposes a localization algorithm based on Evidence K Nearest Neighbor and Grid Clustering algorithm(GC-EKNN).In this algorithm,the local space with positioning requirements is marked as place of interest--ed(POI),algorithm only requires building regional fingerprint database in the POI,reducing the workload of fingerprint collection.In the regional fingerprint database,the grid clustering algorithm is used to divide the grid hierarchy.Then,the region was identified by the Evidence K Nearest Neighbor theory(EKNN),and the target area was determined.Finally,using the grid clustering query neighbor cell within the area range,exploiting the dynamic neighbors screening mechanism from the neighbor cell to select neighbor points,combined with the weighted K Nearest Neighbor algorithm(WKNN)for precision positioning.In this paper,GC-EKNN algorithm module is established on the experimental platform,compared with similar algorithms to verify the performance of GC-EKNN algorithm in positioning efficiency and precision.Experiment showed that the GC-EKNN algorithm makes use of the theory of artificial intelligence EKNN algorithms to decide space separation area,and grid clustering of fingerprints in POI;No matter whether the regional space continuous or not,GC-EKNN algorithm is applicable and can improve the positioning efficiency.At the same time,in this algorithm,using the dynamic neighbors mechanism to select near point for localization.Compared with other similar algorithms,it improves the positioning precision.Finally,the WiFi indoor positioning system based on GC-EKNN is designed and implemented,the feasibility and practical value of the algorithm are verified.