研究一种基于人工神经网络的能区分故障的数字电路测试生成方法，该方法利用电路基本逻辑门的特性和神经网络模型的特点，首先建立测试生成的神经网络模型，然后通过求解网络能量函数的最小值点获得给定类型故障的测试矢量，其研究结果在可区分故障的测试生成方面提供了一种可能的新途径 This paper studies a digital circuit test generation method based on artificial neural network which can distinguish faults from each other. Based on the characteristics of circuit basic logic gates and the characteristics of neural network model, this method first builds the neural network model of test generation, The minimum point to obtain the test vector for a given type of fault provides a possible new way of distinguishing fault-test generation