•  
  •  
 

Keywords

Signal detection, Deep learning, SVM classifier, feature extraction, QPSK, 8PSK and QAM

Document Type

Research Paper

Abstract

During the communication system development over the years, detecting and identifying the signal from the affected noise had the major role and the most attention from the researchers. The deep learning algorithm has become a very attractive tool for distinguishing between signal and noise. Learning and training are the two important steps in designing any deep learning system. The proposed system depends on the support vector machine (SVM). The SVM is one of the most popular learning algorithms in different fields, such as signal processing, image processing, communication, and pattern recognition. The SVM classifier is a supervised learning algorithm that uses the closest data points as "support vectors" to build a hyperplane that divides classes. SVM is used to identify the large RADIOML 2018.01A dataset with various signal schemes. The paper strongly emphasized extracting the dataset's most essential features, which improved Support Vector Machines’ capacity to detect signals in noisy and complicated situations. The measured accuracy for the SVM classier for QPSK, 8PSK, and 16 QAM equals 99.7%, 99%, and 99.6%, respectively. The final measured results show the proposed detection system's correctness, robustness, and flexibility based on utilizing the Support Vector Machines (SVM) classifier; this classifier approves its efficacy in signal detection.

References

Q. V. Pham, N. T. Nguyen, T. Huynh-The, L. B. Le, K. Lee, and W. J. Hwang, Intelligent Radio Signal Processing: A Survey, IEEE Access, 9 (2021) 83818-83850. https://doi.org/10.1109/ACCESS.2021.3087136 Q. Zheng, X. Tian, Z. Yu, Y. Ding, A. Elhanashi, S. Saponara, and K. Kpalma, MobileRaT: A Lightweight Radio Transformer Method for Automatic Modulation Classification in Drone Communication Systems, Drones, 7 (2023) 596. https://doi.org/10.3390/drones7100596 H. Sun, Y. Zhang, F. Wang, J. Zhang and S. Shi, SVM Aided Signal Detection in Generalized Spatial Modulation VLC System, in IEEE Access, 9 (2021) 80360-80372. https://doi.org/10.1109/ACCESS.2021.3084823 A. Linares, B. Mejia, A. Sanchez and G. Kemper, An SVM-based Intelligible Signal Presence Detection Algorithm for FM Signals Demodulated via SDR, 2022 11th Int. Conf., Commun. Circuits Syst., (ICCCAS), Singapore, 2022, 90-95. https://doi.org/10.1109/ICCCAS55266.2022.9823981 M. Du, P. Zhong, X. Cai, D. Bi, DNCNet: Deep Radar , Signal Denoising and Recognition, IEEE Trans. Aerosp. Electron. Syst., 58 (2022) 3549-3562. https://doi.org/10.1109/TAES.2022.3153756 H. Yang, H. Xu, Y. Shi, Y. Zhang, and S. Zhao, A Few-Shot Automatic Modulation Classification Method Based on Temporal Singular Spectrum Graph and Meta-Learning, Appl. Sci., 13 (2023) 9858. https://doi.org/10.3390/app13179858 X. Zhang, T. Li, P. Gong, R. Liu, and X. Zha, Modulation recognition of communication signals based on multimodal feature fusion, Sensors, 22 (2022) 6539. https://doi.org/10.3390/s22176539 S. H. Kim, J. W. Kim, V. S. Doan, and D. S. Kim, Lightweight Deep Learning Model for Automatic Modulation Classification in Cognitive Radio Networks, IEEE Access, 8 (2020) 197532-197541. https://doi.org/10.1109/ACCESS.2020.3033989 S. H. Kim, J. W. Kim, W. P. Nwadiugwu, and D. S. Kim, Deep Learning-Based Robust Automatic Modulation Classification for Cognitive Radio Networks, IEEE Access, 9 (2021) 92386-92393. https://doi.org/10.1109/ACCESS.2021.3091421 T. Huynh-The, C. H. Hua, J. W. Kim, S. H. Kim, and D. S. Kim, Exploiting a Low-Cost CNN with Skip Connection for Robust Automatic Modulation Classification, 2020 IEEE Wireless Commun. Networking Conf., (WCNC), 2020, 1-6. https://doi.org/10.1109/WCNC45663.2020.9120667 S. H. Kim, C. B. Moon, J. W. Kim, and D. S. Kim, A Hybrid Deep Learning Model for Automatic Modulation Classification, IEEE Wireless Commun. Lett., 11 (2021) 313-317. https://doi.org/10.1109/LWC.2021.3126821 T. Huynh-The, C. H. Hua, V. S. Doan, and D. S. Kim, Accurate Modulation Classification with Reusable-Feature Convolutional Neural Network, 2021 IEEE Eighth Int. Conf. Commun. Electron., 2021, 12-17. https://doi.org/10.1109/ICCE48956.2021.9352042 T. J. O'Shea, T. Roy, and T. C. Clancy, Over-the-Air Deep Learning Based Radio Signal Classification, IEEE J. Sel. Top. Signal Process., 12 (2018) 168-179. https://doi.org/10.1109/JSTSP.2018.2797022 Y. Tian , D. Xu, Endong Tong, R. Sun, K. Chen and Y. Li, Toward Learning Model-Agnostic Explanations for Deep Learning-Based Signal Modulation Classifiers,  IEEE Trans. Reliab., 73 (2024) 1529-1543. https://doi.org/10.1109/TR.2024.3367780 X. He, Z. Cao, P. Ji, L. Gu, Sh. Wei and B. Fan, Eliminating the Fading Noise in Distributed Acoustic Sensing Data, IEEE Trans. Geosci. Remote Sens., 61 (2023) 5906510. https://doi.org/10.1109/TGRS.2023.3263159 S. Dong, C. Dong, Z. Li and M. Ge, Gaussian Noise Removal Method Based on Empirical Wavelet Transform and Hypothesis Testing, 2022 3rd Int. Conf. Big Data Artif. Intell. Internet Things Eng., (ICBAIE), Xi’an, China, 2022, 24-27. https://doi.org/10.1109/ICBAIE56435.2022.9985814 J. H. Tyler, M. M. K. Fadul, D. R. Reising and F. I. Kandah, An Analysis of Signal Energy Impacts and Threats to Deep Learning Based SEI, ICC 2022 - IEEE Int. Conf. Commun. Seoul, Korea, Republic , 2022, 2077-2083. https://doi.org/10.1109/ICC45855.2022.9838884 N. Shajihan, Classification of stages of Diabetic Retinopathy using Deep Learning, Bournemouth University United Kingdom, 2020. https://doi.org/10.13140/RG.2.2.10503.62883 David , Evaluation: from precision, recall and F-measure to ROC informedness, markedness and correlation, J. Mach. Learn. Technol., 2 (2011) 37-63. https://doi.org/10.48550/arXiv.2010.16061. V. Clerico, J. González-López, G. Agam, J. Grajal, LSTM Framework for Classification of Radar and Communications Signals, 2023 IEEE Radar Conference (RadarConf23), San Antonio, TX, USA, 2023, 1-6. https://doi.org/10.1109/RadarConf2351548.2023.10149618

Highlights

This study performed signal detection for various types of modulated signals Simulations used pre-processing steps like fading noise, Gaussian noise, low-pass filtering for real-world scenarios Features like energy, power, skewness, kurtosis, zero-crossing rates, phase differences were statistically extracted Applying features to an SVM classifier achieved 98.72% detection accuracy across various noisy signal types

DOI

10.30684/etj.2024.152557.1796

First Page

149

Last Page

158

Download

Share

COinS