Artificial neural network based multiple fault diagnosis in digital circuits

The paper describes a technique, based on the use of Artificial Neural Networks (ANNs), for the diagnosis of multiple faults in digital circuits. The technique utilises different quantities of randomly selected circuit test data derived from a fault truth table, which is constructed by inserting random single stuck-at faults in the circuit. The paper describes the diagnostic procedure using the technique, the ANN architecture and results obtained with example circuits. Our results demonstrate that when the test data selection procedure is guided by test vectors of the circuit a compact, efficient and flexible ANN architecture is achieved.