Parametric faults detection in analog circuits using polynomial coefficients in NN learning

The paper presents an algorithm for parametric fault diagnosis of nonlinear analog circuits. A power supply current waveform IDD is used as an indicator of a device feature. A test signal is filtered using the discrete wavelet transformation, treated as a filter bank, to obtain a component of signal sensitive to changes of device parameters. Coefficients of the polynomial approximating the component are calculated and used to formulate a learning vector of a feedforward neural network. Thus, it is possible to achieve data compression without the considerable loss of information about the tested device. An illustrative numerical example is presented.