一种基于电磁旁路分析的硬件木马检测方法

为探究利用电磁辐射旁路信号检测集成电路芯片中硬件木马的可行性，分析了芯片电磁旁路信号的组成，构建了信号泄漏模型。在阐释霍特林（K-L）变换原理及特点的基础上，提出了利用K-L变换对芯片电磁辐射旁路信号进行信号特征提取的方法，分析含硬件木马芯片（木马芯片）与不含硬件木马芯片（原始芯片）对应特征信号的差异来检测芯片中是否含有硬件木马。通过在针对基于FPGA密码芯片中植入硬件木马并进行对比检测实验的结果表明，利用上述方法能有效分辨出木马芯片与原始芯片所泄漏电磁信号间的差异，达到检测出芯片中硬件木马的目的。

Hardware Trojan detection method based on electromagnetic side-channel analysis

To do research for the feasibility of using Electromagnetic（EM） emanation side-channel to detect hardware Trojan in IC chips, the structure of EM side-channel signal of chip is analyzed and the leaked model about signal is designed. With explaining the principle of Karhunen-Loève（K-L） transform, a method that uses K-L transform to obtain characteristic signal of EM emanation side-channel is introduced. It detects hardware Trojan by analyzing the difference between the characteristic signal of chip with and without Trojan. Experiments of detecting hardware Trojan in FPGA cipher chip show that this method can distinguish effectively the difference between the EM signals of Trojan chip and Genuine chip, and then the Trojan in chip can be detected.