WLAN Indoor Positioning Algorithm Based On PLS

With the rapid development of information technology,Location-based Services(LBS)bring more conveniences to clients.Nowadays,Global Positioning System(GPS)is the most common ways to obtain location information in outdoor environments.These positioning ways can provide higher positioning accuracy for users outdoor,but there are various obstacles which limit the full use of these methods in the field of indoor positioning.Therefore,various positioning technologies for complex indoor environment are proposed.Indoor positioning technology based on received signal strength indicator(RSSI)of WLAN receives more and more attention,because it needs no special hardware and can make full use of the infrastructures equipped with WLAN technology which widely deployed.This article focuses on WLAN-base positioning algorithms,and the main achievements are as follows:The thesis studies the transmission characteristics of wireless signals and the traditional indoor positioning algorithms,and fully analyzes and compares the advantages and disadvantages of each algorithm.In addition,various factors affecting signal transmission are summarized.There are many variables in the RSSI of the WLAN signal and physical coordinates.Especially when the datasets have a large number of independent variables with multi-collinearity,and the number of measured data is not enough,the traditional positioning algorithms will lead to greater errors in positioning accuracy.The thesis presents a novel nonlinear Partial Least Square(PLS)method to address the problem of low precision in WLAN location,the proposed method integrates an inner Relevant Vector Machine(RVM)function with an external PLS framework.PLS is applied to extract the features of the fingerprint database and physical coordinate to reduce the number of the variable dimensions and eliminate the correlations of RSSI.The obtained score matrices are used as the input and output of RVM with an external PLS framework.The coordinates of test points are regressed and predicted by the RVM-PLS algorithm.Simulation experiments are carried out to discuss the parameters affecting the positioning effect?such as the RVM kernel parameters,fingerprint density and the number of APs.The effects of RVM kernel parameters,fingerprint density and the number of APs on the effectiveness of the algorithm model are also discussed.The experimental comparison verify the validity of the PLS algorithm in extracting internal features and the advantage of RVM to establish internal feature mapping.At the same time,compared with the other algorithms,the experimental results show that the proposed algorithm has higher positioning accuracy.The thesis presents another indoor positioning method which based on the Channel State Information(CSI).Compared to RSSI,CSI provides more abundant data.Each set of CSI data is given in the form of a complex matrix which contains the amplitude and phase information of signals on different channels.And the experimental results show that the indoor positioning method based on CSI has higher positioning accuracy,the indoor positioning method based on CSI is relatively simple,and can realizes device-free passive localization.