基于优化型支持向量机算法的硬件木马检测  

Hardware trojan detection based on the optimized SVM algorithm

作　　者：张磊[1] 殷梦婕 肖超恩[1] 董有恒 Zhang Lei;Yin Mengjie;Xiao Chaoen;Dong Youheng(Department of Electronic Information Engineering,Beijing Electronics Science&technology Institute,Beijing 100071,China)

机构地区：[1]北京电子科技学院电子与信息工程系,北京100071

出　　处：《电子技术应用》2018年第11期17-20,共4页Application of Electronic Technique

摘　　要：目前基于侧信道的分析方法是硬件木马检测的主要研究方向。为提高侧信道分析时的准确率和检测速度,提出了基于优化型支持向量机的分类方法。首先利用主成分分析技术对功耗数据进行处理,降低数据特征相关性;然后通过遗传算法提高惩罚系数和核函数参数的选择;最后进行硬件木马分类器的构建。实验结果表明,优化型SVM方法提高了硬件木马分类器的检测速度和准确率,可以有效检测出面积为0. 1%的硬件木马,准确率最高可以提升15.6%,时间消耗减少98.1%。The main research direction of hardware trojan detection is based on the side channel analysis.This paper proposes an optimized support vector machine(SVM)classification method to improve the accuracy and speed when analyzing the data for detection hardware trojan.Firstly,the principal component analysis is used for data dimension reduction.Then,genetic algorithm is used to accelerate the selection of optimal penalty coefficient and kernel function parameters.Finally,the hardware trojan detection classifier is built with SVM.Experimental results show that the optimized SVM method improves the detection speed and accuracy,and can effectively detect hardware trojans with the area of 0.1%,the accuracy can be increased by 15.6%,and the time consumption can be reduced by 98.1%.

关 键 词：硬件木马 支持向量机 主成分分析 遗传算法 木马分类器 

分 类 号：TN918[电子电信—通信与信息系统]