Research On RFID Based Perceptual Recognition Technology

Internet of things technology(IOT)is the use of network communication technology to associate things with things,and realize the perception and recognition of objects.With the rapid development of perceptual recognition technology and the popularization of intelligent devices,more and more researchers begin to study the behavior monitoring,object location and tracking in indoor environment.RFID has the advantages of high precision and low cost.Researchers began to use RFID equipment to realize non-contact perceptual identification system,so as to facilitate human life.However,the information collected by this system contains a lot of noise,which seriously affects the ability of RFID perception and recognition.There is an urgent need to study RFID technology in order to achieve high-precision perceptual recognitionIn this thesis,RFID equipment is used to collect the original signal,preprocess it and extract appropriate features to obtain robust and accurate scene perception and recognition.Firstly,this thesis summarizes the current research status of RFID at home and abroad.Secondly,it introduces the components and working principle of RFID equipment,introduces the classic sensing technology,and compares it with RFID.Thirdly,aiming at two typical perceptual recognition scenarios,the article proposes the RFID based commodity quantity monitoring system and the RFID based liquid classification detection system.In the first system,in order to reduce the use of RFID tags in the self-service supermarket,a method of monitoring the quantity of goods based on RFID is designed.The standard deviation is used to distinguish static samples from dynamic samples.For static samples,the frequency distribution histogram is used to extract static features For dynamic samples,the fast Fourier transform method is used to extract dynamic features.Finally,the classification is calculated according to the KNN Method to identify the quantity of goods.In the second system,in order to improve the efficiency of liquid classification and detection in security inspection,logistics and other fields,the RFID based liquid classification and detection method is designed.Firstly,the collected data is preprocessed based on phase unwrapping,Butterworth low-pass filter and sliding window,and then the corresponding characteristics are extracted through frequency distribution histogram.Finally,the k-means algorithm is used to identify the type of liquid.Experiments verify the feasibility of the system.