Research On Radio Frequency Fingerprint Extraction And Access Control For ZigBee Devices

With the rapid development of Io T technology,the access control for wireless devices has attracted much interest.When facing a large quantity of unsupervised edge devices in the Io T network,the traditional authentication schemes based on MAC address and digital certificate reveal some drawbacks on security.The research on radio frequency fingerprint(RFF)is conducted in this thesis,which is an emerging extraction and identification technology in the field of physical layer security.This thesis explores a safe and reliable authentication scheme for Io T devices.As an inherent hardware characteristic of devices,RFF could uniquely represent the identity of a device and effectively resist the threat from other identity-forged devices.ZigBee devices are selected as the main research objects for new RFF algorithms.On the basis of current research,this thesis focuses on data-independence and anti-noise performance of RFF.By analyzing the characteristics of IEEE 802.15.4 protocol,which ZigBee physical layer complies,the signal preprocessing algorithm before RFF extraction is optimized.Two methods that can extract RFF from arbitrary data segments are proposed.Besides,sufficient experiments are carried out to verify the accuracy,stability and data-independence.Furthermore,this thesis proposes a general Io T access control framework on RFF.The framework uses the proposed method and is applicable to many kinds of Io T devices by normalizing different RFF into a unified form.Finally,an access control system entity for ZigBee is well designed and built.The main work of this thesis is as follows:(1)Aiming to enhance the stability of fingerprint extraction from ZigBee devices,synchronization tracking and symbol-by-symbol compensation is proposed as new data preprocessing algorithms.The synchronization tracking algorithm solves deviation and omission of sampling point caused by crystal oscillator fluctuation.The symbol-by-symbol compensation algorithm solves the problem of unstable frequency offset.These steps make it possible for extracting stable RFF from arbitrary data segments.(2)An RFF extraction method based on cross-power spectrum is proposed.By utilizing the cyclic shift characteristics of ZigBee signals,the cross-correlation power spectrum between received symbols and reference signals is calculated to form a stable and unified RFF feature.The method extends the range of valid data sources from the preamble to entire signal frames compared to previous methods.On the other hand,the method could effectively resist Gaussian noise and performs well in low signal-to-noise ratio(SNR)scenarios.Only 150 spread spectrum chips are needed to identify 50 ZigBee devices with an accuracy rate over 97%.(3)An RFF extraction method based on adaptive filter is proposed.This method uses an adaptive filter to describe the RFF generating process when signals are sent from transmitter to wireless channel.It is proved well data-independent and can be used in different types of communication modem.Using LMS adaptive filter coefficients as RF fingerprints,94% identification accuracy can be achieved on 50 ZigBee devices.(4)A general Io T access control framework based on RFF is proposed.Considering the various types and weak computing ability of Io T devices,this framework details the interaction mechanism of three-party authentication,uniforms the RFF expression format and designs a frame structure which is compatible with IEEE 802.1X protocol.(5)A ZigBee access control system based on RFF is implemented.The system adopts the dataindependent fingerprint methods proposed in this thesis,and control ZigBee access by matching the RFF and MAC address.The system is equipped with a network gateway and a display platform to realize the real-time identity supervision on ZigBee devices in the network.