论文部分内容阅读
为了解决制造变异和噪声对已有硬件木马检测方法的挑战和干扰,提出了一种新的微弱木马信号检测技术,能够在较大的制造变异和噪声的背景下提取出木马特征信号.首先,将木马检测问题建模为特征提取模型,然后提出了一个基于时域约束估计器和主成分投影的统一子空间木马检测方法.并通过特定的子空间投影或重构信号分析,证实弱小的木马信号可以与各种噪声和干扰区分开来.该方法为已有的硬件木马检测方法提供了一种通用的消除制造变异和噪声影响的方法.设计实现了2个时序硬件木马,在ISCAS89基准电路上进行了仿真实验验证,并在FPGA上进行了硬件实物验证,实验结果均表明了所提方法的有效性和高检测精度.
In order to solve the challenge and interference caused by manufacturing variation and noise on the existing hardware Trojan detection methods, a new weak Trojan signal detection technology is proposed, which can extract the characteristics of the Trojan signal in the context of a large manufacturing variation and noise.Firstly, The Trojan detection problem is modeled as the feature extraction model, and then a unified subspace Trojan detection method based on the time-domain constrained estimator and the principal component projection is proposed. Through the analysis of the specific subspace projection or reconstruction signal, it is proved that the weak Trojan The signal can be distinguished from all kinds of noise and interference.This method provides a common method to eliminate manufacturing variation and noise impact for the existing hardware Trojan detection method.It is designed to realize two timing hardware Trojans, On the simulation test verification, and carried out in the FPGA hardware in-kind verification, the experimental results show that the proposed method is effective and high detection accuracy.