基于小波变换去噪的相关功耗分析攻击研究与实现  

Research and implementation of correlation power analysis based on wavelet transform

作　　者：段晓毅[1] 陈东 高献伟[1] 范晓红[1] 靳济方[1] Duan Xiaoyi;Chen Dong;Gao Xianwei;Fan Xiaohong;Jin Jifang(Dept.of Electronic Information Engineering,Beijing Electronics Science&Technology Institute,Beijing 100070,China)

机构地区：[1]北京电子科技学院电子信息工程系,北京100070

出　　处：《计算机应用研究》2020年第4期1119-1124,共6页Application Research of Computers

基　　金：国家自然科学基金资助项目(61701008);中央高校基本科研业务费资助项目(2017LG05,328201801)。

摘　　要：为了提高功耗分析攻击效率,减少噪声影响,研究了小波变换去噪对功耗攻击的影响以及相关功耗分析(CPA)攻击的相关系数与攻击效果的关系,提出使用平移不变量小波法与小波模极大值法对功耗曲线进行去噪预处理。该方法使用卡尔曼滤波法、小波模极大值法与平移不变量小波法对功耗曲线进行去噪预处理,再对原始数据及去噪后数据分别进行CPA。实验结果显示,与原始数据相比,使用平移不变量小波法改进的CPA相关系数比仅使用CPA提高了165%,比卡尔曼滤波法提高了31.4%,比小波模极大值法提高了26.4%,同时攻击成功所需功耗曲线减少了92%。实验结果表明使用平移不变量小波法改进的相关功耗分析攻击效果最好。In order to improve the efficiency of power analysis attack and reduce the influence of noise,this paper studied the influence of wavelet transform denoising on power attack and the correlation coefficient between correlation power analysis and attack effect.For the purpose of denoising the power consumption curve variable,this paper used wavelet method and the wavelet modulus maximum method.The method included the Kalman filter method,the wavelet modulus maximum method and the translation invariant wavelet method.Then the denoised original data were respectively subjected to CPA.Simulation experiments show that compared with the original data,the improved CPA correlation coefficient using the translation invariant wavelet method is 165% higher than using original CPA,31.4% higher than the Kalman filter method,26.4% higher than the wavelet modulus maximum method.Also,the power curve required for successful attack is reduced by 92%.The experimental results show that the improved power analysis using the translation invariant wavelet method has the best effect.

关 键 词：相关功耗分析 平移不变量小波法 小波模极大值法 密码芯片 

分 类 号：TP309.2[自动化与计算机技术—计算机系统结构]