We propose a novel method that applies least-square support vector machines (LS-SVM) to denoising of impulse response signal for aircraft flight flutter test. This method is based on time series prediction using LS-SVM. Since the signal to noise ratio (SNR) varies with amplitude for the decaying property of damped sinusoid, the beginning data points with high SNR is used for training and prediction of the subsequent data with low SNR. In order to improve the performance of denoising, singular value decomposition (SVD) filtering is employed for signal preprocessing. Finally, the simulations and experiment on real flight test data demonstrate effectiveness and efficiency of our approach.
Denoising of impulse response using LS-SVM and SVD for aircraft flight flutter test
2006-01-01
391244 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Flutter flight test of the RANGER 2000 aircraft
TIBKAT | 1995
|Flutter Prediction from Flight Flutter Test Data
AIAA | 2001
|