An integrated fault pattern recognition method of satellite control system is put forward, of which the key components are feature extraction and support vector machine. In order to work out the most effective feature for fault detection, kernel principal component analysis is adopted to build the accurate principal component model which can not only simplify feature extraction but also compress the feature space dimensions of nonlinear data. Multi classification support vector machine is constructed to recognize the fault pattern, of which the parameters are optimized by adopting particle swarm optimization algorithm. The simulation results indicate the method is capable of detecting fault and recognizing fault pattern well and truly in time.
An integrated fault pattern recognition method of satellite control system using kernel principal component analysis and support vector machine
2014-08-01
190498 byte
Conference paper
Electronic Resource
English
Face Recognition Using Modular Principal Component Analysis and Support Vector Machines
British Library Conference Proceedings | 2006
|British Library Online Contents | 2013
|