WLAN Indoor Positioning Algorithm Based On KPCA-RBF Neural Network

With the rapid development of computer technology and communication technology,people pay more attention to location-based services(LBS).LBS have broad application prospects in commercial use,personal affairs and other fields.Depending on different environments,the positioning system can be divided into indoor positioning system and outdoor positioning system.Global positioning system(GPS)and cellular positioning system are popular outdoor positioning system.these two kinds of systems can provide excellent outdoor positioning service,but the positioning results cannot be accepted when used indoors.With the popularization and application of wireless local area networks(WLAN),most public places have deployed WLAN.It is the opportunity that WLAN-based in-door positioning technology becomes feasible.Because WLAN-based position?ing system can fully utilize existed wireless network infrastructures with accu-rate positioning results,this factor makes WLAN-based positioning technology is superior to others.this article focuses on WLAN-base positioning algorithms,and the main achievements are as follows:Analyze transmission properties of the wireless signal and the traditional WLAN-based indoor positioning algorithms.Compare the advantages and dis-advantages of positioning algorithms.Propose a positioning algorithm based on kernel principal component analysis(KPCA)and radial basis function neural network(RBF NN).This algorithm firstly uses KPCA to abstract the positioning features of RSS(Received Signal Strength)signals which are sampled on fingerprint points,then uses RBF NN to establish a nonlinear relationship between the positioning features and position coordi-nates.With this nonlinear relationship,we can finally predict the test points'coordinates.Analyze several kinds of signal preprocessing algorithms.Use these algorithms to preprocess the RSS signals,and then use RBF NN to establish a nonlinear relation-ship between the positioning features and position coordinates.Test points co-ordinates can be predicted by this nonlinear relationship.Finally,analyze the performance of algorithms by contrast experiments.Experiment results show that,KPCA-RBF NN positioning algorithm provides a better positioning accuracy than traditional positioning algorithms.