基于智能代理者和神经网络的数字电路故障诊断

Fault diagnosis in digital circuit based on aaent and neural network

介绍了一种对板级数字电路多个故障同时进行智能诊断的新方法。该法融合了智能代理和神经网络两种技术的优点,前者是根据专家经验,通过对电路板自动测试的结果进行分析实现电路故障的快速定位,而后者则利用被测电路所对应的约束电路神经网络模型快速实现故障的完备测试集。该法在板级数字电路的多故障并行诊断方面具有较高的智能性和实用性。

In this paper, a novel intelligent method for fault diagnosis in digital circuit board is presented. It combines the advantage of Agent technique and ANN technique. The former one is-used for rapidly rough locating of fault ,according to expert＇s experience, through analyzing the test pattern, and the latter can acquire test sets for fault detection through self-contained neural networks models of circuit. This approach can complete the parallel diagnosis of multiple-fault, also it has great intelligence and practicability.