Home / Current Issue / Paper 1706640
Deep Learning-Based Spectrum Management to Enhance the Performance of Cognitive Radio Network Using MobileNet
Subject area: Science,Engineering and Technology · Area of research: Wireless Communication, Deep learning
Abstract
Cognitive radio (CR) is a leading-edge technology in fifth-generation (5G) network. CR network (CRN) performance can be augmented by effective implementation of spectrum management, which is a decisive function. Signal classification plays a critical role in enhancing spectrum management. Deep learning-based spectrum management (DLSM) is a transformative technology to enhance the performance of CRN. The present work proposes a DLSM using a predefined convolutional neural network (CNN) architecture, MobileNet. The proposed DLSM was developed using a dataset with 1000 constellation diagrams of several digital modulation schemes at a signal to noise ratio (SNR) = 10 dB. The dataset was divided in to 60% for training, 20% for validation, and 20% for testing. In the proposed DL model, the dataset is pre-processed, and feature extraction is conducted using convolution layers; then, classification of images is accomplished using fully connected layers. The results outperformed with 89.4% accuracy, 90% precision, 89% recall, and 89% F1 score. The proposed DLSM exhibits substantial waveform classification performance; hence, it can be recommended for spectrum management in CRN. The dataset was generated and the work implemented using the open-source Python programming language and the licensed Colab Pro platform.
Keywords
Cognitive radio network, deep learning- based spectrum sensing, CNN, accuracy, confusion matrix, AUC.
How to cite this paper
@article{1706640,
author = {M. V. S. Sairam, Raju Egala, Hanumanthu Rajasekhar, Killada Sai Nohith},
title = {Deep Learning-Based Spectrum Management to Enhance the Performance of Cognitive Radio Network Using MobileNet},
journal = {Iconic Research And Engineering Journals},
year = {2024},
volume = {8},
number = {6},
pages = {274-279},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1706640.pdf},
abstract = {Cognitive radio (CR) is a leading-edge technology in fifth-generation (5G) network. CR network (CRN) performance can be augmented by effective implementation of spectrum management, which is a decisive function. Signal classification plays a critical role in enhancing spectrum management. Deep learning-based spectrum management (DLSM) is a transformative technology to enhance the performance of CRN. The present work proposes a DLSM using a predefined convolutional neural network (CNN) architecture, MobileNet. The proposed DLSM was developed using a dataset with 1000 constellation diagrams of several digital modulation schemes at a signal to noise ratio (SNR) = 10 dB. The dataset was divided in to 60% for training, 20% for validation, and 20% for testing. In the proposed DL model, the dataset is pre-processed, and feature extraction is conducted using convolution layers; then, classification of images is accomplished using fully connected layers. The results outperformed with 89.4% accuracy, 90% precision, 89% recall, and 89% F1 score. The proposed DLSM exhibits substantial waveform classification performance; hence, it can be recommended for spectrum management in CRN. The dataset was generated and the work implemented using the open-source Python programming language and the licensed Colab Pro platform.},
keywords = {Cognitive radio network, deep learning- based spectrum sensing, CNN, accuracy, confusion matrix, AUC.},
month = {December},
}