Research On Influencing Factors Of Radio Frequency Fingerprint Subtitle

With the development of information attack and defense technology,encryption algorithms and security pro-tocols that rely on cracking costs are gradually being challenged.Equipment identification based on radiofrequency fingerprints is an important technology for a new generation of network security systems.According to previous experiments,radiofrequency fingerprints will be affected by various physical factors such as channel environment,device startup time,temperature,etc.,thereby reducing the accuracy of device identification.Studying the influence of physical factors on radiofrequency fingerprints can improve the accuracy of device identification.This thesis selects 802.11 n wireless devices as the research object,analyzes the OFDM PSDU frame structure,and introduces several existing radiofrequency fingerprint extraction methods.Generalized linear regression and other methods are used to quantify the influence of physical factors on radiofrequency fingerprints.The work of this thesis is concluded as follows.· The OFDM PSDU frame structure is studied,the existing carrier frequency offset extraction methods are intro-duced,and a 6-dimensional transient feature extraction method is proposed.This thesis uses 40 devices from 4manufacturers for experiments.Experimental results show that all the three characters can improve classification accuracy.· In view of the influence of startup time on frequency offset,this thesis determines the optimal regression model according to mathematical characteristics of data set,and determines the optimal training duration according to evaluation methods.The wireless device identification system based on frequency offset preprocessing is established.The cleaned data set is used for equipment classification experiments and a classification accuracy of 94.2% is carried out,which is higher than the classification accuracy before cleaning.· In view of the influence of environmental temperature on frequency offset,this thesis focuses on the changes of radiofrequency fingerprints at room temperature.The reasons for the differences of the fitting coefficients are analyzed.A temperature standardization method is proposed to further improve the classification accuracy of the equipment.