基于模式识别和自适应滤波的功耗攻击技术

Performing Power Analysis with Pattern Recognition and Adaptive Filtering

引入复杂的随机噪声是一种有效的抗功耗攻击技术。虽然复杂功耗噪声产生模块具有多种不同的工作状态，但每种工作状态都具有相对稳定的功耗行为。文中将这种功耗行为称为功耗模式。本文提出应用模式聚类和识别技术来确定功耗攻击时各样本中噪声所属何种模式，并且应用自适应滤波技术消除功耗样本中的噪声。实验结果表明，这种预处理方法有效地改善了功耗攻击的效果。

Additive noise is a kind of typical power analysis resistant implementing technique. The complex noise generator has more than one kind of power consumption, each kind of power consumption can be considered as a pattern. Based on the combination of pattern recognition and adaptive filtering, the noise in each sample can be eliminated. The experimental result shows that the presented pre-processing method can enhance the ability of power analysis attack greatly.