Potentiality of Data Fusion in Analog Circuit Fault Diagnosis

Features extracted from single domain information cannot maximally reveal the state of the circuit since parametric fault features of analog circuits are quantified according to the application of the circuit. To combat this shortcoming, features from time, frequency, wavelet and statistical domain have been fused to construct ultimate fault features of the circuit. Data fusion has been performed in two steps, data whitening and Principal component analysis (PCA). The fused features are used to train SVM classifier for fault diagnosis of analog circuits. The proposed method is illustrated with the example of Sallen-Key band pass filter circuit and four OpAmp biquad high pass filter circuit. The accuracy of fault classification of the proposed method with fused features is found considerably higher than that with individual domain features.