论文部分内容阅读
针对数字电路中多侵略线串扰时滞故障将使电路不能正常工作的问题,提出了一种有效的测试生成算法。该算法首先使用Kernighan-Lin-Fiduccia-Mattheyses(KLFM)方法在受害点处把电路分为左右2个部分,然后对右半部分电路采用神经网络算法把故障效应传播到输出端,最后对左半部分电路根据受害点时滞故障的方程计算出受害点的最大串扰时滞故障测试生成矢量。在ISCAS’85和ISCAS’89国际标准电路上的实验结果表明本算法具有较高的故障覆盖率和较短的测试生成时间。
Aiming at the problem that the multi-invasion crosstalk delay fault in digital circuit will make the circuit not work normally, an effective test generation algorithm is proposed. The algorithm first uses Kernighan-Lin-Fiduccia-Mattheyses (KLFM) method to divide the circuit into two parts at the victim point, and then uses the neural network algorithm to propagate the fault effect to the output on the right half of the circuit. Finally, Some circuits calculate the maximum crosstalk delay fault test generation vector of the victim according to the equation of the fault time-delay fault. The experimental results on ISCAS’85 and ISCAS’89 international standard circuits show that this algorithm has higher fault coverage and shorter test generation time.