Study On 3D Indoor Positioning Based On Fingerprinting And Sensors Of Smartphone

With the rapid development of information society,people's demand for location-based services is expanding.In outdoor environments,satellite navigation and positioning system has been relatively mature,and can meet the demand of the real-time accurate positioning,and indoors,because of the complexity of the indoor environment,a general positioning system that can provide reliable and efficient indoor location services has not appeared until now.The finger-printing location technology based on WiFi signal is the focus of industry researchers because of low cost,easy to achieve and wide coverage.However,due to the method based on received signal strength(RSS),complexity and variability of the indoor environment will be larger influence on the RSSI,therefore,it is hard to achieve continuous,stable,the high precision positioning effect simply based on WiFi Signal to realize positioning.For positioning technology based on pedestrian dead reckoning(PDR)using smart phone,its accuracy is higher in a short time,but it is a kind of relative positioning and long-time error accumulation effect is very obvious.How to combine the advantages of different positioning technologies to realize a low cost,high precision and high performance fusion positioning system is one of the main research contents of this paper.This paper focuses on few key problems such as long time to obtain location,poor precision,the existing positioning blind spots and inaccurate floor identification,achieves the purpose of improving the positioning accuracy,real-time performance and realizing the multi-floor continuous and stable positioning effect,adopts the fusion theory based on WiFi and mobile phones of inertial sensor positioning technology.The main research contents of this paper as follows:(1)Against the problems of real-time positioning and accuracy of location fingerprinting,based on the idea of clustering analysis,a kind of indoor positioning method combined improved fuzzy kernel clustering(KFCM)with the weighted K-neighbor(WKNN)is put forward,aimed at shortening the time of online positioning and improve the positioning accuracy.By using clustering by fast search and find of density peaks(CFSFDP)to determine the clustering number and initial cluster centers,overcome traditional KFCM algorithm to select the dependence of the initial clustering center and lead to instability of clustering results,on this basis,adopt WKNN matching method to improve the positioning accuracy.The experiment shows that the method proposed in this paper can effectively reduce the location calculation and time in the premise of ensuring certain accuracy compared with the indoor positioning method without clustering.(2)A hybrid positioning system using Kalman filter based on WiFi location fingerprint positioning technology and smartphone PDR positioning technology is proposed for WiFi fingerprint positioning accuracy is not ideal and locate result has large fluctuations.In the positioning system,in order to promote the fusion positioning effect,use a detection method "rise with zero point + peak detection + dynamic threshold" for steps detection,and use the magnetometer,gyroscope and gain movement direction angle,to reduce the positioning error of PDR.The experiment shows that the result of fusion positioning has been improved in positioning accuracy and stability.(3)A floor identification method based on coarse classification and KFCM clustering center is proposed.First,for the problem of dimension for fingerprint database is higher,use the method of sub database to store the fingerprint data,and establish the database of the wireless Access Point(AP)distribution interval between different floor,when the floor identification,use rough classification judgment at first,and then according to the clustering center distance to judge if the result is not the only.The experimental results show that this method can achieve a floor identification accuracy of 97.8% when the number of AP deployed is enough.