基于相关性分析的硬件木马检测方法  

Hardware Trojan Detection Method Based on Correlation Analysis

作　　者：刘长龙[1] 赵毅强[1] 史亚峰[1] 冯紫竹[1] 

机构地区：[1]天津大学电子信息工程学院,天津300072

出　　处：《计算机工程》2013年第9期183-185,195,共4页Computer Engineering

基　　金：天津市自然科学基金资助重点项目(12JCZDJC20500)

摘　　要：为提高硬件木马检测的准确率,提出一种基于相关性分析的检测方法。在完成木马功耗建模的基础上,提出并分析利用经典相关系数进行木马检测的可行性以及存在的缺点,根据木马检测的特点,优化检测系数,给出利用区间重叠比作为木马判定依据的检测方法。实验结果表明,与采用经典相关系数的方法相比,该方法在降低约6%检测准确率的前提下,能使鲁棒性提高1倍以上。To improve the accuracy of the hardware trojan detection, a method based on the correlation analysis is presented. After a power modeling of the trojan, the feasibility and defect of the detection method by using the classical correlation coefficient are analyzed, the detection coefficient is optimized according to the characteristics of trojan detection, and the detection method depended on the region overlap ratio is presented. Experimental results show that this method can provide more than two times of robustness when the detection accuracy declined about 6% compared with the classical correlation coefficient.

关 键 词：硬件木马 功耗建模 功耗分析 相关性分析 木马检测 

分 类 号：TP407[自动化与计算机技术]