Fault Diagnosis Method of Spacecraft Control Systems Based on PCA-ResNet

Abstract Due to the complex space environment, spacecraft telemetry signals are accompanied by a large amount of noise, and accuracy fault diagnosis is low directly using original signals. This paper proposes method for control systems based on principal component analysis (PCA) residual network (ResNet). Firstly, grayscale images generated denoising signal system through PCA; Secondly, input into extract deep-level features; Finally, Softmax classifier used classification realize system. The research results show that diagnostic proposed in this reaches 95.33%, which higher than other models, can be actual systems.