FastICA在功耗分析去噪中的应用

The application of the FastICA denoising in power analysis

提出了一种应用独立成分分析技术提高功耗分析中信号信噪比的方法。搭建了侧信道分析平台,采集了功耗数据,并对功耗数据进行了基于负熵的快速固定点算法的去噪实验,信噪比提升19 dB。最后提出了通过更换g函数实现优化的方式,对不同g函数在功耗分析中的去噪能力进行了对比。结果表明,更换后的去噪能力比之前提高4 dB以上。

This paper proposed a novel signal processing method to improve the signal-noise ratio（SNR） of power analysis based on independent component analysis（ICA）.In this experiment, fast fixed-point algorithm（FastICA） was employed to de-noise the collected power data with the established side channel analysis platform, and the experimental result showed a 19 dB increment of SNR. Finally, an optimization method by changing g function was proposed, and de-noise capabilities were analyzed. The result demonstrated that the SNR of the algorithm with changed g function increased by more than 4 dB compared with the previous one.