Researches On Target Locating And Tracking Of Indoor Multi-floor

Location based service(LBS)application requirements have received widespread attention with the popularity and widespread use of mobile devices,and have penetrated into many aspects of modern life.In recent years,localization algorithms based on received signal strength have been widely studied and applied in indoor positioning.Indoor positioning technology is roughly divided into special equipment-based positioning technology,rangingbased positioning technology and location fingerprint-based positioning technology.At present,most domestic and foreign research and applications use location fingerprint positioning technology to achieve indoor positioning.Location fingerprint-based positioning technology has the advantages of no additional infrastructure,low cost and better positioning accuracy.Further,the value of RSS(Received Signal Strength)obtaining by WiFi is easier than obtaining other physical features such as AOA(Angle of Arrival)and TOA(Time of arrival).Compared with the distance positioning,fingerprint positioning technology can cope with indoor complex and changeable environment more easily.Therefore,WiFi-based fingerprint positioning has been widely studied and applied in indoor positioning.At present,most indoor positioning studies are based on two-dimensional space,that is,single-story research.For multi-floor research and identification,two-dimensional space research cannot meet the demand.At present,the indoor environment is mostly multi-floor scenes(such as shopping malls,airports,hospitals,etc.),so the indoor positioning of multiple floors has gradually attracted the attention of scholars.Aiming at the problems of low accuracy,low stability and high complexity of the existing multi-floor positioning technology,this paper proposes a multi-floor positioning and tracking method based on linear discriminant analysis(LDA),which can improve the stability of indoor positioning while reducing the computational complexity,and can get higher floor discrimination accuracy and positioning accuracy.The main research work of this paper includes:1)A multi-floor identification system based on WiFi is proposed for multi-floor identification.The floor discrimination of the system is divided into an offline phase and an online phase.In the offline phase,meshing is performed and data is collected in each grid to establish a complete offline WiFi location fingerprint database.Then,a multi-floor discrimination algorithm based on multi-class linear discriminant analysis is used to establish a floor discrimination model.In this algorithm,only a small amount of WiFi fingerprint information is required to obtain the floor number of the target,and the accuracy is high.In the online phase,the floor identification model and the voting mechanism are combined to determine the final floor number of the target,and the improved K-nearest neighbor(KNN)algorithm is designed to obtain the location information of the target node.2)Aiming at the problem that single technology or method is difficult to guarantee tracking and positioning accuracy,this paper proposes an improved particle filter tracking and localization algorithm based on Pedestrian Dead Reckoning(PDR).After determining the floor number and initial position information of the target node,it uses sensors which built-in smartphone to collect relevant information of the target node(such as direction information,acceleration information,pressure information,etc.)and determine the motion state of the target,establishes a motion model of the indoor target node,and then combines the model to improve the state equation and the observation equation of the particle filter.At last,track the motion trajectory of the target node and improve the accuracy of the target tracking.The experimental results show that the system can get a more accurate floor number in a multi-floor environment and get a more accurate tracking of target movement.