A Novel Sensing Strategy through Denoising Autoencoder and Ensembling Methods | IEEE Conference Publication | IEEE Xplore

A Novel Sensing Strategy through Denoising Autoencoder and Ensembling Methods


Abstract:

Cooperative sensing is suitable for sensing the primary user (PU) activity in the cognitive radio network. However, the problem with cooperative sensing is false sensing ...Show More

Abstract:

Cooperative sensing is suitable for sensing the primary user (PU) activity in the cognitive radio network. However, the problem with cooperative sensing is false sensing users (FSUs). The FSUs tackled in this paper are the Yes False Sensing (YFS) and No False Sensing (NFS) users, which consistently report high and low energy statistics to the decision authority. The sensing reliability is improved by denoising and reducing the effects of abnormalities in sensing due FSUs' participation with the denoising autoencoder. The denoised soft energy reports and then forwarded to the ensemble classifier. Classification performance of the PU activity based on the cleaned dataset is investigated for the decision tree, k-nearest neighbor, neural network and ensemble classification method, and random forest classifier, where ensemble classification with AdaBoost ensembling method offers better results.
Date of Conference: 19-21 October 2022
Date Added to IEEE Xplore: 25 November 2022
ISBN Information:

ISSN Information:

Conference Location: Jeju Island, Korea, Republic of

Funding Agency:


References

References is not available for this document.