Research On Indoor Positioning Method Based On RFID Technology

Radio Frequency Identification(RFID)is a contactless system that automatically identifies specific targets and obtains relevant information based on Radio Frequency signals.With the development of society,people have a growing demand for locationbased services.Underground garage navigation and library book positioning all need accurate location information.RFID technology is widely used in the field of indoor positioning due to its small size and low price.Due to the influence of indoor obstacles,the noise interference and multi-path effect of the signal in the indoor propagation will cause certain positioning errors.The existing indoor positioning technology based on RFID has some shortcomings,such as low positioning accuracy and great environmental impact.In this paper,aiming at improving the positioning accuracy and the robustness of the positioning system,indoor positioning technology based on RFID is studied.The main work is as follows:(1)This thesis analyzes and studies the composition and working principle of RFID-based your positioning system,and analyzes the current commonly used RFID positioning system from the aspects of cost,system complexity,positioning effect,etc.In view of the characteristics of interference caused by the complex and multiple changes of indoor environment,gaussian filter is adopted to preprocess the collected signals,eliminate some invalid data,and reduce the interference caused by indoor noise and multipath effect.(2)For traditional RFID localization method based on path loss model of positioning accuracy is not high,locate the amount for a long time,are greatly influenced by environmental problems,this paper puts forward a kind of indoor location algorithm based on dual neural network model,establish the BP network and the network within DNN dual neural network model,the BP network output path loss coefficient of n and the RSSI value input to the network within DNN positioning results,with the change of indoor environment and take the corresponding path loss coefficient,effectively solve the error caused by indoor environment changes.At the same time,gaussian filter is used to preprocess the collected received signal strength,which reduces the influence of some invalid data on the result.The algorithm effectively improves the positioning accuracy and the robustness of the system.(3)For gaussian filter can only noise data,not smooth output data and error on the positioning problem,adopted Gaussian-Kalman filter for data preprocessing to improve positioning accuracy,at the same time in order to solve the traditional indoor locating method cannot locate tracking is proposed a fusion of RFID localization algorithm of computer vision.The method uses computer vision positioning to assist RFID positioning,the computer vision positioning to detect the target output pedestrian foot coordinates,RFID positioning output label coordinates,the fusion algorithm is used to match the results of the two,so as to output the location information of a specific person carrying a label.