A New Method for Satellite Control System Fault Pattern Recognition combining Multi-Classification SVM with Kernel Principal Component Analysis

A new method for satellite control system fault pattern recognition is put forward, the most important components of which is feature extraction. Kernel principal component analysis is adopted to work out the most effective feature for fault detection by building a accurate principal component model, and this model makes the feature extraction simple and declines the dimensions of feature space. Furthermore in order to recognize the fault pattern a multi classification support vector machine (SVM) is worked out, and the particle swarm optimization algorithm is used to optimize the parameters of it. The results of simulation indicate that this method can not only detect the fault but also recognize the fault pattern.