Research On User Identification And Authentication Based On RFID Wireless Sensing

Advances in ambient intelligence and ubiquitous computing are driving the unprecedented development of smart space,and the premise of smart space to personalize services for specific users is the identification of the user who interacts with it.User identification and authentication are of great significance to enhance the personalization and security of smart space.Currently,the widely used user authentication methods are based on users' physiological biometrics(such as faces,fingerprints,etc.),but this type of user identification method that relies on physiological biometrics requires active user participation and suffers from spoofing attacks.For example,wearing a facial mask may spoof face recognition systems.In fact,in addition to physiological biometrics,many human daily motions show distinct characteristics which are often more difficult to mimic and can be used as behavioral biometrics for user identification.As a result,the reliability of physiological biometric based user identification systems can be further enhanced by adding a new dimension of behavioral modality.As one of the indispensable infrastructures of Internet of Things(Io T),Radio Frequency Identification(RFID)technology has been widely used in various environments,providing sensing abilities in addition to traditional object identification,such as indoor localization,vital sign monitoring,and activity recognition.Passive RFID tags,because of their simple structure,low cost,and flexible coverage,can be easily attached to the surface of an object to build or expand an intelligent environment,which also conforms to the concept of integration of environment and computing proposed by ubiquitous computing.This paper explores the feasibility of deploying an RFID tag array for contactless user identification.Firstly,RFID tags are arranged in a spatially distributed tag array,and then the low-level channel characteristics collected from the RFID backscattered signal by a commercial RFID reader are used to capture the behavior characteristics of users,enabling user identification in a more natural manner to reduce the user's active participation.The main research work of this paper is as follows:(1)Deploy an RFID tag array composed of multiple tags to capture behavioral characteristics embodied in human motions for user identification and authentication.In order to make full use of the spatial-temporal information contained in the tag array,this paper first represents the non-uniform data points from multiple tags reported by the reader as a multi-modal sensing data stream,and then designs multi-modal neural networks to effectively fuse fine-grained information in time,space and modal dimensions from the RFID tag array.(2)Propose a behavioral biometric based user identification and authentication scheme through the motions captured from sequential human activities using RFID,Au-Id.This paper designs a sequence labeling based motion segmentation method,and leverages the temporal correlation of human daily activities to segment the target motion from continuous activities for user identification.In order to perform user identification,this paper develops a convolutional neural network to extract high-level features that effectively represent the user's physical and behavioral characteristics,and reduces the training cost when a new user joins through transfer learning.At the same time,user authentication is achieved through one-class classifiers.(3)Conduct a series of experiments to evaluate the effectiveness of the proposed contactless user identification and authentication scheme based on RFID wireless sensing,proving the effectiveness of the scheme for target motion segmentation,user identification and user authentication.This thesis first introduces the background and significance of user identification and authentication research,then reviews the related work,and describes the detailed design of the user identification and authentication scheme Au-Id.Finally,a comprehensive experiments were carried out to evaluate the performance of Au-Id.The experimental results show that AuId can effectively segment the target motion from human daily activities and achieve accurate user identification and authentication,which is of great significance to the further development of smart space.