Research And Application Of 3D Positioning Method Based On Indoor WiFi

With the development of society,more and more high-rise buildings appear in cities,but GPS(Global Positioning System)can not be used to locate indoors.This has caused great difficulties to the location management of indoor personnel,and some public places have a large flow of people.If people are not effectively managed,it will pose a threat to public safety.Therefore,how to accurately determine the threedimensional position of users in the indoor is an urgent problem to be solved.Based on WiFi location fingerprint indoor positioning technology,this paper studies and improves WiFi location fingerprint indoor positioning technology by combining theoretical analysis with experimental simulation.The main work of this paper is as follows:Firstly,the existing indoor positioning technology is introduced and analyzed,and the advantages and disadvantages of various indoor positioning technologies are compared.The reasons for choosing WiFi-based location fingerprint positioning technology are indicated.Secondly,aiming at the disadvantage of low reliability of position fingerprint database in location fingerprint positioning technology,the data filtering method is optimized.By counting the strength values of multiple sets of signals collected from the same signal transmitter at the same sampling point,and setting threshold according to the signal distribution,the data in different thresholds are optimized according to different methods,which improves the reliability of position fingerprint database.Thirdly,aiming at the disadvantage of poor real-time performance and large amount of calculation of traditional positioning methods,the hierarchical clustering method is used to reduce the spatial dimension.By clustering the sampling points of all floors,the Euclidean spatial distance is calculated by using the signals collected from the locating points and the centroid signals of each cluster after clustering,and the floor number of the cluster where the shortest Euclidean spatial distance is located is taken as a pair.The preliminary identification of target location reduces the amount of data calculation;aiming at the shortcoming of low accuracy,the weighted K-nearest neighbor algorithm is improved.Through the adaptive method,the K-value selection is more scientific,and the algorithm is used to match in the reduced-dimension area to improve the positioning accuracy.Finally,the effectiveness of the improved method is verified by field data acquisition,simulation experiments and comparison with other methods.