.png){kind=icon}
.png){kind=icon}
Return
CBiLSTMDSNet for Automatic Modulation Classification in 5G and Beyond
.png){kind=icon}
PRE
AI
DOI：10.1029/2024RS008131.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Wireless networks in 5G and Beyond 5G (B5G) will be more dynamic and heterogeneous, necessitating the use of multi-strand wave shape. The biggest serious obstacle in such a large-scale system, exclusively in non-cooperative situations, is determining the modulation type to recognize the precise encoding type employed by the broadcaster at the current moment in order to successfully decode the information. Typical modulation classification requires expert signal processing algorithms that perform noise reduction and estimation of signal parameters, namely, carrier frequency and signal power. Hence, this research introduces a revolutionary Automatic Modulation Classification (AMC) model that is built on reconfigurable Convolutional BiLSTMDSnet (CbiLSTMDSnet). The proposed reconfigured deep learning architecture is technologically advanced by merging the convolutional neural network, BiLSTM, and DSnet. A-MAX pooling layer is also included in the proposed model to boost classification accuracy. Furthermore, the Dimensionality reduction was also carried out with the help of a Restricted Boltzmann Machine to reduce the training time. Finally, a dropout layer and a Gaussian noise layer are added to the proposed neural network model to minimize the signal noises for effective modulation classification. The simulation results in evidence that the proposed AMC model outperforms existing classification models with improved accuracy of 99.8% in less time.
Keywords:
convolution neural network (CNN)
Bi-directional Long Short-Term Memory (BiLSTM)
DenSenet (DSnet)
automatic modulation classification (AMC)
5G wireless communication
restricted Boltzmann machine (RBM)
dimensionality reduction
Journal
R
IF:
1.5
Papers:
64
Citations:
5.0K