A Novel Approach for Diagnosis of Analog Circuit Fault by Using GMKL-SVM and PSO

Abstract

This paper presents a novel analog circuit fault diagnosis approach using generalized multiple kernel learning-support vector machine (GMKL-SVM) method and particle swarm optimization (PSO) algorithm. First, the wavelet coefficients' energies of impulse responses are generated as features. Then, a diagnosis model is constructed by using GMKL-SVM method based on features. Meanwhile, the PSO algorithm yields parameters for the GMKL-SVM method. Sallen-Key bandpass filter and two-stage four-op-amp biquad lowpass filter fault diagnosis simulations are given to demonstrate the proposed diagnose procedure, and the comparison simulations reveal that the proposed approach has higher diagnosis precision than the referenced methods.