CNN-Based RF Fingerprinting Method for Securing Passive Keyless Entry and Start System

作     者：Hyeon Park SeoYeon Kim Seok Min Ko TaeGuen Kim 

作者机构：Department of Smart Convergence SecuritySoonchunhyang UniversityAsanKorea Department of Information Security EngineeringSoonchunhyang UniversityAsanKorea 

出 版 物：《Computers, Materials & Continua》 (计算机、材料和连续体（英文）)

年 卷 期：2023年第76卷第8期

页      面：1891-1909页

核心收录：

学科分类：08[工学] 081104[工学-模式识别与智能系统] 0805[工学-材料科学与工程（可授工学、理学学位）] 0811[工学-控制科学与工程] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：supported by Institute of Information&communications Technology Planning&Evaluation(IITP)grant funded by the Korea Government(MIST)(No.2022-0-01022 Development of Collection and Integrated Analysis Methods of Automotive Inter/Intra System Artifacts through Construction of Event-Based Experimental System). 

主　　题：RF fingerprint cepstral coefficient convolutional neural network 

摘      要：The rapid growth of modern vehicles with advanced technologies requires strong security to ensure customer safety.One key system that needs protection is the passive key entry system(PKES).To prevent attacks aimed at defeating the PKES,we propose a novel radio frequency(RF)fingerprinting method.Our method extracts the cepstral coefficient feature as a fingerprint of a radio frequency signal.This feature is then analyzed using a convolutional neural network(CNN)for device identification.In evaluation,we conducted experiments to determine the effectiveness of different cepstral coefficient features and the convolutional neural network-based model.Our experimental results revealed that the Gammatone Frequency Cepstral Coefficient(GFCC)was the most compelling feature compared to Mel-Frequency Cepstral Coefficient(MFCC),Inverse Mel-Frequency Cepstral Coefficient(IMFCC),Linear-Frequency Cepstral Coefficient(LFCC),and Bark-Frequency Cepstral Coefficient(BFCC).Additionally,we experimented with evaluating the effectiveness of our method in comparison to existing approaches that are similar to ours.