Research On PDR+CSI Fingerprint Indoor Location Technology

In recent years,with the increasing demand for Location Based Services(LBS),many positioning systems have emerged.The outdoor environment relies on the Global Navigation Satellite System(GNSS)to achieve centimeter-level high-precision positioning.However,in indoor environments,GNSS satellite signals are not available,so the development of indoor positioning technology has become a research hotspot.Considering the limitations of Pedestrian Dead Reckoning(PDR)and Received Signal Strength(RSS),two commonly used indoor positioning technologies,the human behavior-assisted PDR localization algorithm and the channel state information of WiFi are studied.Passive fingerprint localization algorithm based on Channel State Information(CSI).Based on this,a PDR+CSI fingerprint indoor positioning method is proposed.The specific research contents and conclusions are summarized as follows:(1)The traditional PDR positioning algorithm ignores the influence of human motion mode on the positioning result.When the human motion mode changes,if the traditional PDR algorithm is used,the plane positioning accuracy will be reduced a lot,and the pedestrian height change is not considered.Therefore,the human body is proposed in this paper.Motion mode assisted PDR positioning algorithm.Support vector machine(SVM)is used to classify human motion patterns.Statistical features are used to select the best features for classifier training through Recursive Feature Elimination(REF),and different sliding window size classification results are discussed.The effectiveness of the human motion pattern recognition assistant PDR algorithm to improve indoor positioning accuracy is illustrated by experiments.(2)The RSS information of WiFi is only the superposition of various information of WiFi,which is greatly affected by the surrounding environment,resulting in the fingerprint positioning accuracy of WiFi RSS is not high,and the CSI information has finer-grained channel information,adopting CSI amplitude and The phase information is used to establish the fingerprint database.The KNN classification algorithm is used for CSI passive fingerprint location.The RSS information,CSI amplitude information,CSI amplitude + phase information are used as the location accuracy of the location fingerprint.The proposed CSI-based amplitude and phase are illustrated.The validity of the KNN classification passive fingerprint localization algorithm for information.(3)In the indoor positioning of PDR,the initial position needs to be determined,and the positioning error of the PDR algorithm will become larger and larger with time,and the KNN classification passive fingerprint positioning based on CSI amplitude and phase information has uniform accuracy.The characteristics of the positioning method combine the two positioning methods to locate,which indicates that the PDR+CSI fingerprint joint positioning accuracy is better than the single PDR positioning accuracy.