非对称分布差错模型下特征分析器混叠概率研究

Research on Aliasing Probability for Signature Analyzer Using Asymmetric Error Model

本文讨论了使用非对称分布差错模型准确计算特征分析器混叠概率的一种新颖的方法,通过对混叠过程的分析,构造了由两个相同的特征分析器组成的一个虚拟特征分析器,其中一个特征分析器被参考电路的输出序列驱动,另一个被有故障的被测电路的输出序列驱动,根据所得到虚拟特征分析器对应的马尔可夫链的状态概率分布,就可以准确计算对于任意测试输入序列长度的特征分析器的混叠概率。研究了线性和非线性特征分析器的混叠概率的动态特性。分析了故障激活概率对混叠概率的重要影响。与独立分布差错模型相比,非对称分布差错模型在混叠概率的建模、准确计算和分析中能提供更全面的信息。最后,给出了应用实例,验证了文中方法的可行性和有效性。

This paper discusses a novel approach to calculate the exact aliasing probability for signature analyzer, which tracks the states of a fictitious signature analyzer comprised of two identical signature analyzers, one is driven by the output sequence of the fault free reference circuit and the other is driven by the output sequence of the faulty circuit under test. According to the state transition probability distribution of Markov Chain corresponding to the fictitious signature analyzer, we can calculate the exact aliasing probability for any test input vector length. Fur- thermore, we investigate the dynamic properties of aliasing probability for signature analyzer and aliasing probability for nonlinear compactor. The activation probability associated with single stuck - at fault is extremely important in determining the aliasing probability in signature analyzer. The analysis in the paper indicates that the asymmetric error model will provide more detailed information in modeling, analysis and exact calculation of aliasing probabili- ty. Finally, simulation results show the effectiveness of both the analysis and the approach proposed in the paper.