Research On The Key Technology Of Fingerprint Database Using Non-autonomous AP Deployment In WLAN Indoor Fingerprint Positioning

With the advent of the mobile Internet era,a large number of services need to know the user terminal where the physical location.Outdoor positioning can rely on GPS or Beidou system,but based on the global positioning system of GPS positioning users cannot solve the problem of indoor positioning accuracy.Wireless Local Area Network(WLAN)widely used in indoor and popularized smart phones,the corresponding WLAN indoor positioning technology began to rise,where the location of fingerprint location method is the current research hot spots and has good application prospects of positioning technology.WLAN location Fingerprint indoor location technology relies on the correlation of received signal strength(RSS)and spatial location,and building a compact and effective location fingerprint library is the key to realize reliable location.However,in the environment of nonautonomous deployment of Access Points(APs),there are some problems such as random noise,terminal differences and excessive APs in the location fingerprint database.In order to solve the above problems of fingerprint library and improve positioning accuracy and stability,this paper has done the following work:Firstly,in order to solve the problem of random noise interference to fingerprint,this paper analyzes all kinds of fingerprint original sample processing algorithms,and then compares the characteristics of each method,then proposes an improved algorithm based on RSS outlier filtering algorithm-The abnormal value filtering algorithm,and the actual scene data test.Experimental results show that the improved outlier filtering algorithm can effectively reduce the impact of random noise on the fingerprint and improve the positioning accuracy and positioning stability.Secondly,we focus on the influence of terminal differences on positioning accuracy.We investigate various existing solutions,and compare the performance of the algorithms with the simulation and experiment.The experimental results show that the cosine method and the spatial mean normalization have a good effect on resolving the terminal difference,and both the localization accuracy and localization stability are improved greatly.Thirdly,we focus on the study of AP selection algorithm for fingerprint location in case of too many APs,and propose a new AP selection algorithm based on regional frequency.The experimental results show that the AP algorithm based on regional frequency is better than other AP selection algorithms in locating performance,which can effectively solve the problem that the redundant AP leads to the increase of positioning complexity and improve the efficiency of the algorithm.Fourthly,presents a design and implementation of a fingerprint database for positioning system in a non-autonomous deployment AP environment.The localization system integrates the improved outlier filtering algorithm,SMN,AP selection algorithm based on regional frequency,using the number of non-fixed AP storage design and two stages of the processing flow.The prototype system based on our proposed method can provide reference for the research and implementation of the existing indoor fingerprint location system.