Research On Device Identification And Group Key Generation Method Based On Radio Frequency Fingerprinting

With the increasing number of internet of things(Io T)devices,more and more important information is carried in wireless communication networks.Due to the increased computing power of attackers and the new requirements for energy consumption in Io T scenarios,it is increasingly difficult for traditional security systems based on computing security to adapt to the security requirements of future wireless communication networks.The Physical layer security(PLS)technology technology that uses channel fingerprinting or radio frequency fingerprinting(RFF)to achieve security functions has been paid more and more attention by researchers.This thesis revolves around RFF,explores the generation mechanism of RFF,the generation mechanism of RFF was explored,the identification method of RFF was proposed to solve the problem of open set device identification,the method is verified by Universal Software Radio Peripheral(USRP),and a solution is given to the problem of group key generation in multi-source scenarios by using RFF.Firstly,this thesis studies the generation mechanism of RFF.Details the distortion that may be generated in the process of transmitting baseband signals from wireless devices,and establishes a mathematical model of the generation mechanism of RFF.In this thesis,the practical application scenarios of wireless communication are analyzed,and a signal transmission model including RFF is given considering the wireless channel.On the basis of the model,combined with the RFF generation mechanism,the signals sent by the device under different RF distortions are simulated.Simulation results show that the signals from devices with different RF distortions still contain their unique characteristics after passing through random channels.Therefore,it is feasible to identify wireless communication devices based on RFF.Secondly,for the problem of radio frequency fingerprinting identification,this thesis proposes an intelligent identification scheme of RFF based on the convolutional neural network framework.The identification methods for closed set identification and open set identification problems are presented,and the experimental verification based on USRP is completed.Simulation experiments are carried out on different modulation schemes,and it is found that the accuracy of the sample processing scheme based on time-frequency analysis is about 10% higher than that based on the time-domain original signal.Tests are carried out in variable modulation recognition,and it is proved that both recognition schemes are effective,but the recognition accuracy drops by about 20%.On the basis of the theoretical analysis of the RFF recognition scheme,this thesis further uses the USRP to build the RFF recognition simulation experiment hardware platform.Based on the RF signal data of different devices collected by the platform,the identification algorithm proposed in this thesis is verified.The results show that the RFF identification scheme of wireless devices proposed in this thesis is effective based on the actual data collected by the USRP in this thesis.The radio frequency characteristics of open set devices are unknown.In order to identify open set devices,this thesis proposes an open set device identification scheme based on majority voting and threshold detection.The majority voting scheme enables the recognition algorithm to improve the recognition accuracy by about 5% at the expense of certain computing resources.After verification on the actual data set,it is found that the detection accuracy of unknown devices is greatly increased under the condition of losing a certain accuracy of known devices.Finally,this thesis designs a group key generation scheme based on XOR reflexivity and RFF uniqueness for data security sharing in multi-source user multicast communication scenarios.This solution is the first to add RFF identification to the key generation process,and uses spread spectrum technology and RSA algorithm to solve the problem of multi-source terminal resource selection conflict.The key generation rate,resource selection conflict and key generation efficiency of the group key generation scheme are theoretically analyzed,and the feasibility and effectiveness of the scheme are proved.Considering the group key eavesdropping model,the security of the group key is simulated and analyzed.The simulation results show that,different from the traditional key generation scheme,the security performance of the group key generation scheme is related to the number of users,and the key is more difficult to crack when the number of users is large.