Fault Diagnosis and Health Assessment for Super-Heterodyne Receivers Based on ITD-SVD and LR

As a typical device widely used in electronics and information systems, the super-heterodyne receiver plays a key role in the whole system. This study proposes a method of fault diagnosis and health assessment for super-heterodyne receivers based on intrinsic time-scale decomposition (ITD)-singular value decomposition (SVD) and logistic regression (LR). First, a state observer based on radial basis function (RBF) neural network is designed to calculate the residual error between the actual and estimated signal outputs. Second, proper rotation components of the residual error are obtained by ITD. Then the singular values of the components are extracted by SVD to form feature vectors. Finally, a second RBF neural network is trained by the features to realize the classification of common fault modes, and the LR model is trained to estimate the health state of the super-heterodyne receiver. The feasibility and effectiveness of the proposed scheme are demonstrated by the results of simulation experiments.