改进边界Fisher分析近邻选择的硬件木马检测

针对旁路分析技术对小规模硬件木马检测精度低的问题，提出基于边界Fisher分析的硬件木马检测方法.定义规则式选择近邻样本，以减小样本与其同类近邻样本间距离和增大样本与其异类近邻样本间距离的方式构建投影子空间，在不对数据分布作任何假设的前提下，提取原始功耗旁路信号中的差异特征，实现硬件木马检测. AES加密电路中的硬件木马检测实验表明，该方法能够检测出占原始电路规模0.02%的硬件木马，优于已有的检测方法.

Hardware Trojan detection based on improved marginal Fisher analysis nearest neighbor selection

A hardware Trojan detection method based on marginal Fisher analysis was proposed aiming at the problem of low detection accuracy for small-scale hardware Trojan by side-channel analysis. The rule was defined to select the nearest neighbor samples. Then the projection subspace was constructed by reducing the distance between the samples and their nearest neighbor samples in the same class and increasing the distance between the samples and their nearest neighbor samples in different class. The difference features in the original side-channel signals was found without any assumptions about data distribution, and the detection of hardware Trojan was achieved. The hardware Trojan detection experiment in AES encryption circuit shows that the hardware Trojan whose scale is 0.02% of the original circuit can be detected by the method. The method is better than existing detection methods.