Analog Circuits Based Fault Diagnosis using ANN and SVM

Abstract

In this study, we provide a technique for identifying analog errors using a neural network and an SVM (SVM). The study's major objective is to produce a trustworthy diagnostic based on a technique that reduces testing durations by resolving the problem of component tolerances.The suggested strategy uses an artificial neural network and a backward propagation mechanism. The impact of methods like Principal Component Analysis on feature extraction is discussed in this work. The simulation results show that the technique is effective and efficient for fault identification in tolerant mixed-signal circuits.