Research On RFID Indoor Positioning Algorithm Based On BP Neural Network Error Correction

Driven by the current wave of Internet of things,the near-field communication technology of Internet of things services has become a hot research topic,especially the increasing demand for the use of indoor positioning to achieve personnel management and dynamic control of important items.RF positioning technology has been widely used in business,medical,industrial management and other fields.This technology can realize the precise positioning of the physical location of the object.How to improve the precision of the positioning method,especially in the emergency environment,will be of great significance to the development of personnel management and the Internet of things.RFID positioning technology uses radio frequency to exchange data in non-contact two-way communication to realize the purpose of mobile device identification and positioning.High precision positioning information can be obtained in a few milliseconds,and the transmission range is large and the cost is low.But RFID is greatly affected by the environment.The environmental noise will make the positioning accuracy of RFID greatly reduced,resulting in a large positioning error.Therefore,the main research direction of this paper is to reduce the environmental noise,reduce the positioning error and improve the positioning accuracy.The main research work of this paper is as follows:(1)In this paper,we evaluate the accuracy and analyze the performance of common location algorithm models,including LANDMARC model,Vire model and bvire model.(2)In this paper,RFID positioning error correction technology based on neural network is proposed to reduce the error by learning the error of the environment.The positioning accuracy of the existing algorithm is improved.The error correction algorithm of BP neural network based on bvire model is realized in MATLAB simulation platform and laboratory environment,and it is verified by experiment and simulation.Simulation and experiments show that(1)compared with the algorithm proposed in this paper,the positioning accuracy of bvire model is reduced from the error range of bvire model [0.2-0.45] to [0.1-0.3].(2)The algorithm proposed in this paper is effective for the point correction with large error.