基于核主成分分析的硬件木马检测方法研究

针对旁路信号样本在高维空间中的分布,提出了一种基于核主成分分析的硬件木马检测方法,该方法能够找出旁路信号样本分布中的非线性规律,将高维的旁路信号映射到低维子空间同时更精确地反映旁路信号样本的分布特性,从而发现由木马引起的非线性特征差异。针对AES加密电路植入约占电路3%的组合型木马并进行检测,实验结果表明,该方法能够有效分辨基准电路与含木马电路之间旁路信号的非线性特征差异,实现木马的检测,并取得比K-L变换更好的检测效果。

Hardware Trojan Detection Method Based on Kernel Principal Component Analysis

Aimed at the distribution of side-channel signal sample in high-dimensional space, a Hardware Trojan detection method based on kernel principal component analysis is proposed. This method can identify the nonlinear law in distribution of side-channel signal sample, project high-dimensional side-channel signal onto low-dimensional subspace, and more accurately reflect the distribution characteristics of side-channel signal sample, so that non-linear characteristics of differences caused by Trojan are found. In against the AES encryption circuit,the combined Trojan circuit about 3% is implemented, and detection experiments are performed. Experimental results show that, the method can effectively distinguish the difference about nonlinear characteristics between side-channel signal of the reference chip and side-channel signal of chip with Trojan, achieve the detection about Trojan, and get a better detection result than the K-L Transform.