Diagnostics of incipient faults in analog circuits

Abstract

Diagnosis of incipient faults for analog circuits is very important, yet very difficult. A novel approach for incipient faults in analog circuits is proposed. Firstly, the statistical property feature vector, which is composed of range, mean, standard deviation, skewness, kurtosis, entropy and centroid, is used to reflect the global property of output response. Then, the least squares support vector machine (LSSVM) is used for diagnostics of the incipient faults in analog circuits. Traditionally, multi-fault diagnosis for analog circuits based on SVM usually uses a single feature vector to train all binary SVM classifier. However, in fact, each binary SVM classifier has different classification accuracy for different feature vectors. Therefore, the Mahalanobis distance (MD) based on particle swarm optimization (PSO) is proposed to select a near-optimal feature vector and decrease the dimensions of the feature vector for each binary classifier. The experiment results show as following: (1) The accuracy using the near-optimal feature vectors is better than the accuracy using a single vector; (2) The consuming time of the nearoptimal feature vectors selected by MD based on PSO is reduced about 98% in comparison to the time of the optimal feature vectors selected by the exhaustive method.