On minimal set of test nodes for fault dictionary of analog circuit fault diagnosis
Journal of Electronic Testing: Theory and Applications
Ensembles of Neural Networks for Fault Diagnosis in Analog Circuits
Journal of Electronic Testing: Theory and Applications
Multiple catastrophic fault diagnosis of analog circuits considering the component tolerances
International Journal of Circuit Theory and Applications
Hi-index | 0.01 |
A new neural network-based fault classification strategy for hardmultiple faults in analog circuits is proposed. The magnitude of theharmonics of the Fourier components of the circuit response atdifferent test nodes due to a sinusoidal input signal are firstmeasured or simulated. A selection criterion for determining the bestcomponents that describe the circuit behaviour under fault-free(nominal) and fault situations is presented. An algorithm thatestimates the overlap between different faults in the measurementspace is also introduced. The learning vector quantization neuralnetwork is then effectively trained to classify circuit faults.Performance measures reveal very high classification accuracy in bothtraining and testing stages. Two different examples, whichdemonstrate the proposed strategy, are described.