| The emergence of “human-machine-thing” intelligent perception and recognition technology based on the Internet of Things has gained significant attention as a hot research direction,thanks to the rapid development of technologies like the Internet of Things and artificial intelligence.Out of the numerous technologies used for perception and identification,Radio Frequency Identification(RFID)technology has found extensive utilization in multiple sectors including identity validation,tracking of logistics,and indoor localization.Out of a wide range of perception and recognition technologies,wireless communication-based Radio Frequency Identification(RFID)technology has gained immense popularity in numerous industries including identity verification,logistics monitoring,and indoor positioning,owing to its distinct benefits such as effective data handling and non-line-of-sight communication.This thesis focuses on the RFID technology in the perception and recognition layer of the Internet of Things,and explores the research and implementation of two methods: cross-domain interconnection between mobile terminals and RFID tags,and gesture recognition based on RFID tags,from the perspective of RFID signal characteristics.The primary achievements and research components of this study can be outlined as follows:(1)A method for cross-domain interconnection between mobile terminals and RFID tags is proposed.This thesis designs a pairing algorithm based on a synchronization mechanism and uses user-defined gesture actions to match the phase data of RFID devices with the acceleration sensor data captured by smartphones to achieve pairing of smartphones with RFID tags.To solve the noise pollution problem in RFID data transmission,this thesis uses discrete wavelet denoising processing.The experimental results show that the proposed algorithm is feasible and practical,and has potential application value in the future cross-domain interconnection of mobile terminals and RFID tags.(2)A method for gesture recognition using RFID tags is proposed.This thesis proposes two algorithms: gesture start and end point detection algorithm and DTW-KNN algorithm based on gesture segmentation.The two algorithms extract gesture action data from the phase data collected by RFID reader devices and recognize different gesture actions,respectively.In the process of algorithm implementation,in order to solve the sudden change of phase data,this thesis uses phase compensation technology for phase data processing,and uses SG smoothing filtering to denoise the collected raw data to reduce errors caused by natural reasons such as experimental devices.The RFID gesture recognition proposed in this thesis has better privacy protection,accuracy,and stability than traditional gesture recognition,and is also more portable and flexible,thus having better application prospects.(3)To validate the feasibility and soundness of the proposed approaches,a model system was assembled.In terms of practical implementation,this thesis integrates the design of hardware components,overall architectural plan,function module design,and the arrangement of correlated tables to actualize the authentication of RFID tags and gesture recognition.Furthermore,comprehensive functional evaluations were conducted on the complete system.The successful construction of this prototype system provides strong support for the practical application of the proposed cross-domain interconnection and gesture recognition algorithms in this thesis. |
