Research On Identity Authentication Technology Based On RFID Array

In recent years,with the flourishing of the Internet of Things(Io T)technology,the related technology of environment perception,data transmission and cloud computing have become popular research directions.The main purpose of the perception layer of the Io T is to convert the analog signals of the physical world into digital signals which computers can transmit,store and calculate.The main tasks in the perception layer are the positioning of users and objects,user authentication and actions recognition.Among them,the user authentication part is the key technology of the Io T's perception layer.After the user's identity is authenticated,personalized services can be further provided.This thesis focuses on the use of Radio Frequency Identification(RFID)arrays to achieve identity authentication by sensing user gestures and breathing characteristics.The existing research works has the following problems: 1)In the gesture-based authentication technology,when using Dynamic Time Warping(DTW)algorithm for feature matching,the algorithm complexity is relatively high;when using classification algorithms such as neural networks for user matching,the demand for samples is large,and the training time is long;2)In the breath-based authentication technology,the authentication accuracy is low by using common time-frequency features for user matching.In response to the above two issues,this thesis has done the following work:First of all,in the RFID array based gesture authentication technology,a histogram based low complexity feature extraction algorithm is adopted,which can effectively extract the waveform characteristics of the signal.Compared with the widely used DTW algorithm,it has lower time complexity while ensuring a certain accuracy rate,reducing system delay,and increasing the practicability of the system.Secondly,in the RFID array based breath authentication technology,the signal's Intrinsic Mode Function(IMF)component is obtained by performing Empirical Mode Decomposition(EMD)on the signal.Then from the perspective of time-frequency domain joint analysis,the fluctuation index,EMD energy entropy and Hilbert marginal spectrum characteristics of IMF components are extracted respectively to effectively characterize the user's breathing information,and ultimately improve the system authentication accuracy.Finally,through experimental verification,this thesis confirms that in the RFID array based gesture authentication system,the use of a low complexity histogram based feature extraction algorithm enables the system to obtain a high authentication accuracy.Under the condition of 8 users,the average authentication accuracy rate has reached 98.7%;at the same time,in the RFID array based respiratory authentication system,the features extracted in this thesis can be used to achieve higher authentication accuracy than existing work.Under the condition of 8 users,the average authentication accuracy has reached 99.1%,which confirming the effectiveness of the features extracted in this thesis.