Research On Positioning Accuracy And Algorithms Of Location Fingerprint Indoor Based On WiFi

WiFi-based location fingerprint indoor positioning is the most widely used indoor positioning technology at present.Offline database establishment and online positioning algorithm affect the positioning accuracy of unknown nodes.The relationship between fingerprint density and positioning accuracy and the optimization of online matching algorithm are scientific problems to be solved.In this paper,first of all,through theoretical derivation and computer fitting method,the relationship between the final location of unknown nodes in the offline phase and the geometry composed of N fingerprint points around them is analyzed.On the basis of this relationship,the offline database with different fingerprint point density and different arrangement mode is designed,and the effect of fingerprint point density of offline database on the location accuracy is studied.Then,from unknown nodes Starting from the positioning algorithm,through the analysis of the problem that the(RSSI)value of unknown node is similar to that of the received signal strength indicator RSSI of fingerprint point in the weight k-nearest neighbor(WKNN)algorithm,but the actual distance is far away,an online matching algorithm based on cell is proposed,and the positioning accuracy of the two algorithms under different density and different access point(AP)layout is compared.Finally,it is compiled The main research results of cell matching algorithm program are as follows:1.In the geometric figure composed of N reference nodes with high similarity,the final location error of unknown nodes depends on the size of the graph area of reference nodes,the area of large location area increases,and the location error increases.The location area is related to the distance between reference nodes,that is,the density of fingerprint.Under the algorithm,the location range is small and the positioning accuracy is high.There is a significant linear positive correlation between the test node density and the positioning accuracy of the offline database.The experimental results are consistent with the computer fitting results.Even with the same off-line database with the largest density of location algorithm,the location accuracy of unknown nodes is the highest,and with the decrease of off-line data reference node density,the location accuracy gradually deteriorates.2.The cell-based fingerprint matching algorithm is based on the smallest grid(cell)in the experimental area instead of the traditional WKNN algorithm to compare the similarity of fingerprint points.By computer fitting the relationship between the final location of unknown nodes and the location of cells,it is found that with the increase of the distance between fingerprint points,the actual location of unknown nodes and the calculated Euclidean distance are the most The probability of small cell consistency is more than 85%,which basically meets the requirement that the cell of unknown node is the cell of real location or the cell next to it.Therefore,the cell location algorithm effectively avoids the problem that the Euclidean distance is similar,but the actual unknown node is far away from the fingerprint point.A cell location algorithm program is written in C# language,and the system is tested and analyzed in two ways: different fingerprint density and different AP layout.The results show that the location accuracy of the cell location algorithm is higher than that of the traditional WKNN algorithm,and the quadrangular layout method is due to the difference between each AP.Because of the large distance,less overlap area and less signal interference,the location result is the best.