eduzhai > Applied Sciences > Engineering >

Wavelet Classification for Over-the-Air Non-Orthogonal Waveforms

  • Save

... pages left unread,continue reading

Document pages: 6 pages

Abstract: Non-cooperative communications using non-orthogonal multicarrier signals arechallenging since self-created inter carrier interference (ICI) exists, whichwould prevent successful signal classification. Deep learning (DL) can dealwith the classification task without domain-knowledge at the cost of trainingcomplexity since neural network hyperparameters have to be extensively tuned.Previous work showed that a tremendously trained convolutional neural network(CNN) classifier can efficiently identify feature-diversity dominant signalswhile it failed when feature-similarity dominates. Therefore, a pre-processingstrategy, which can amplify signal feature diversity is of great importance.This work applies single-level wavelet transform to manually extracttime-frequency features from non-orthogonal signals. Composite statisticalfeatures are investigated and the wavelet enabled two-dimensionaltime-frequency feature grid is further simplified into a one-dimensionalfeature vector via proper statistical transform. The dimensionality reducedfeatures are fed to an error-correcting output codes (ECOC) model, consistingof multiple binary support vector machine (SVM) learners, for multiclass signalclassification. Low-cost experiments reveal 100 classification accuracy forfeature-diversity dominant signals and 90 for feature-similarity dominantsignals, which is nearly 28 accuracy improvement when compared with the CNNclassification results.

Please select stars to rate!

         

0 comments Sign in to leave a comment.

    Data loading, please wait...
×