DERİN ÖĞRENME TABANLI MODÜLASYON TANIMA

Year 2023, , 123 - 140, 30.04.2023

Mehmet Merih Leblebici Ali Çalhan Murtaza Cicioğlu

https://doi.org/10.17482/uumfd.1161509

Abstract

Haberleşme teknolojilerinde her geçen gün artan sinyal çeşitliliği, bu sinyallerin tanımlanması ve sınıflandırılması gerekliliğini ortaya çıkarmıştır. Beşinci nesil (fifth generation, 5G) ve ötesi kablosuz haberleşme teknolojileri, birçok uygulama için vazgeçilmez iletişim araçları haline gelmiştir. Otomatik modülasyon tanıma (automatic modulation recognition, AMR) tekniği, özellikle yeni nesil nesnelerin interneti, akıllı şehirler, otonom araçlar ve bilişsel radyo gibi birçok uygulama için temel bileşen haline gelmiştir. Bu çalışmada sekiz farklı modülasyon türü kullanılarak bir veri seti oluşturulmuş ve derin öğrenme (deep learning, DL) algoritmalarından olan evrişimli sinir ağları (convolutional neural network, CNN) kullanılarak farklı sinyal-gürültü oranlarında (signal-to-noise ratio, SNR) modülasyon türü sınıflandırılması yapılmıştır. Sonuç olarak SNR değerleri 10 dB, 20 dB ve 30 dB iken CNN ile sınıflandırma işleminde sırasıyla %80,76, %99,89 ve %100 doğruluk sağlanmıştır.

Keywords

Otomatik Modülasyon Tanıma, Evrişimli Sinir Ağları, Derin Öğrenme, 5G ve Ötesi Haberleşme Teknolojileri

Deep Learning Based Modulation Recognition

Abstract

The increasing signal diversity of communication technologies has revealed the need that these signals to be defined and classified. Fifth-generation (5G) and beyond wireless communication technologies have become indispensable communication tools for many applications. The automatic modulation recognition (AMR) technique has become a key component for many applications, especially the next-generation internet of things, smart cities, autonomous vehicles, and cognitive radio. In this study, a data set was created using eight different modulation types and modulation classification was made at different signal-to-noise ratios (SNR) using convolutional neural networks (CNN) from deep learning (DL) algorithms. As a result, while the SNR values were 10 dB, 20 dB, and 30 dB, CNN provided 80.76%, 99.89%, and 100% accuracy in the classification process, respectively.

Keywords

Automatic Modulation Recognition, Convolutional Neural Networks, Deep Learning, 5G and Beyond Communication Technologies

References

• 1. Ansari, S., Alnajjar, K. A., Saad, M., Abdallah, S. ve El-Moursy, A. A. (2022) Automatic Digital Modulation Recognition Based on Genetic-Algorithm-Optimized Machine Learning Models, IEEE Access, 10, 50265–50277. doi:10.1109/ACCESS.2022.3171909.
• 2. Dulek, B. (2017) Online Hybrid Likelihood Based Modulation Classification Using Multiple Sensors, IEEE Transactions on Wireless Communications, 16(8), 4984–5000. doi: 10.1109/TWC.2017.2704124.
• 3. Hu, L., Jiang, H., Lu, R. ve Liu, C. (2021) Signal Classification in Real-time Based on SDR using Convolutional Neural Network, Proceedings of 2021 IEEE 2nd International Conference on Information Technology, Big Data and Artificial Intelligence, ICIBA2021, (Iciba), 893–898. doi:10.1109/ICIBA52610.2021.9687958
• 4. Lin, S., Zeng, Y., ve Gong, Y. (2022) Learning of Time-Frequency Attention Mechanism for Automatic Modulation Recognition, IEEE Wireless Communications Letters, 11(4), 707–711. doi: 10.1109/LWC.2022.3140828.
• 5. O’Shea T. J., Corgan, J., ve Clancy, T. C. (2016) Convolutional radio modulation recognition networks, Communications in Computer and Information Science, 629, 213–216. doi: 10.1007/978-3-319-44188-7_16.
• 6. Shi F. Y., Hu, Z. M., Yue, C. S. ve Chen, Z. C. (2022a) Combining neural networks for modulation recognition, Digital Signal Processing, 120, 103264. doi: 10.1016/J.DSP.2021.103264.
• 7. Shi F. Y., Yue, C. S. ve Han, C. (2022b) A lightweight and efficient neural network for modulation recognition, Digital Signal Processing, 123, 103444. doi: 10.1016/J.DSP.2022.103444.
• 8. Wang, Y., Liu, M., Yang, J. ve Gui, G. (2019) Data-Driven Deep Learning for Automatic Modulation Recognition in Cognitive Radios, IEEE Transactions on Vehicular Technology, 68(4), 4074–4077. doi:10.1109/TVT.2019.2900460
• 9. Wang Y., Gui, J., Yin, Y., Wang, J., Sun, J., Gui, G., Gacanin, H., Sari, H. ve Adachi, F. (2020) Automatic Modulation Classification for MIMO Systems via Deep Learning and Zero-Forcing Equalization, IEEE Transactions on Vehicular Technology, 69(5), 5688–5692. doi: 10.1109/TVT.2020.2981995.
• 10. Zeng, Y., Zhang, M., Han, F., Gong, Y. ve Zhang, J. (2019) Spectrum Analysis and Convolutional Neural Network for Automatic Modulation Recognition, IEEE Wireless Communications Letters, 8(3), 929–932. doi:10.1109/LWC.2019.2900247
• 11. Zhang F., Luo, C., Xu, J., Luo, Y. ve Zheng, F. (2022) Deep learning based automatic modulation recognition: Models, datasets, and challenges, Digital Signal Processing, 129, 103650. doi: 10.1016/J.DSP.2022.103650.
• 12. Zhou, R., Liu, F. ve Gravelle, C. W. (2020) Deep Learning for Modulation Recognition: A Survey with a Demonstration IEEE Access, 8, 67366–67376. doi:10.1109/ACCESS.2020.2986330