Prediction of Power Amplifier Performance via Fine and Coarse Modeling Along with Deep Neural Network

Öz

The power amplifier (PA) is a nonlinear design for which an accurate characterization is required for modeling and optimizing effectively. To tackle this difficulty, we present a method based on the fine and coarse modeling approach along with the implementation of deep neural networks (DNNs). For this case, firstly the executed transistor is modeled with the X-parameters and the DNN, as the ’fine modeling’. Then, the S-parameters are modeled with the help of configured hidden-layer structure at the previous step as the ’coarse modeling’ leads to facilitate the overall PA sizing. Finally, the PA is modeled through the optimized DNN, which leads to estimating the performances of PA at the extended frequency in terms of S-parameters, output power, power gain, and efficiency. The presented fine and coarse modeling is powerful enough to configure the hidden-layer configuration of DNNs without any need for other optimization methods for determining the number of hidden layers with neurons in each one. The presented methodology is validated by designing and optimizing a PA with a power gain of more than 11 dB and a power-added efficiency of around 60% operating with 600 MHz band frequency.

Anahtar Kelimeler

• fine and coarse modeling
• deep neural network
• optimization
• power amplifier

Derin Sinir Ağı Tabanlı İnce ve Kaba Modelleme Yoluyla Güç Kuvvetlendirici Performansının Tahmini

Öz

Güç kuvvetlendiricileri (GK), tasarımında yüksek doğruluklu bir karakterizasyonun kritik öneme sahip olduğu lineer olmayan bir devre bloğudur ve zorlayıcı isterlere göre tasarlanması için etkin bir şekilde modellenip optimize edilmesi gerekmektedir. Bu amaçla, öncelikle yapıda kullanılan transistörün X parametreleri ve DNN kullanılarak "ince modeli" elde edilir. Ardından, önceki adımda yapılandırılmış gizli katman yapısıyla transistörün bu sefer S parametreleri elde edilir, çünkü bu "kaba model" genel PA boyutlandırmasını kolaylaştırmaktadır. Son olarak, GK, optimize edilmiş DNN aracılığıyla modellenir ve bu da GK’nın genişletilmiş frekanstaki performanslarının S parametreleri, çıkış gücü, güç kazancı ve verimlilik açısından tahmin edilmesine olanak tanır. Önerilen ince ve kaba modelleme yöntemi, DNN’lerin gizli katman yapılandırmasını belirlemek için yeterli olup, gizli katman sayısı veya her katmandaki nöron sayısı gibi hiperparametreleri belirlemek için ek bir optimizasyon yöntemine ihtiyaç duymamaktadır. Sunulan yöntem, 600 MHz bant frekansında çalışan, 11 dB’den fazla güç kazancı ve yaklaşık %60’lık güç eklenen verimliliğe sahip bir GK’nin tasarlanması ile doğrulanmıştır.

Anahtar Kelimeler

• ince ve kaba modelleme
• derin sinir ağı
• optimizasyon
• güç kuvvetlendirici

Kaynakça

1. W. Li, H. Zeng, L. Huang, et al., “A review of terahertz solid-state electronic/optoelectronic devices and communication systems,” Chinese Journal of Electronics, vol. 34, no. 1, pp. 26–48, 2025. DOI: 10.23919/cje.2023.00.282.
2. U. Ghafoor, “Strengthening data confidentiality in 6g cognitive networks through secrecy rate optimization,” IEEE Communications Letters, pp. 1–1, 2025. DOI: 10.1109/LCOMM.2025.3537285.
3. D. Zeng, H. Zhu, G. Shen, et al., “A compact 24–30-ghz gan front-end module with coupled-resonator-based transmit/receive switch for 5g millimeter-wave applications,” IEEE Transactions on Microwave Theory and Techniques, pp. 1–13, 2025. DOI: 10.1109/TMTT.2025.3530435.
4. D. Wang, M. Aziz, M. Helaoui, and F. M. Ghannouchi, “Augmented real-valued time-delay neural network for compensation of distortions and impairments in wireless transmitters,” IEEE Transactions on Neural Networks and Learning Systems, vol. 30, no. 1, pp. 242–254, 2019. DOI: 10.1109/TNNLS.2018.2838039.
5. J. Wang, R. Su, J. Lv, G. Xu, and T. Liu, “Signal reconstruction deep residual neural network-based bandwidth augmented methods for dpd linearization,” IEEE Microwave and Wireless Technology Letters, vol. 33, no. 3, pp. 243–246, 2023. DOI: 10.1109/LMWC.2022.3217691.
6. Y. Wu, A. Li, M. Beikmirza, et al., “Mp-dpd: Low-complexity mixed-precision neural networks for energy-efficient digital predistortion of wideband power amplifiers,” IEEE Microwave and Wireless Technology Letters, vol. 34, no. 6, pp. 817–820, 2024. DOI: 10.1109/LMWT.2024.3386330.
7. Z. Liu, X. Hu, L. Xu, W. Wang, and F. M. Ghannouchi, “Low computational complexity digital predistortion based on convolutional neural network for wideband power amplifiers,” IEEE Transactions on Circuits and Systems II: Express Briefs, vol. 69, no. 3, pp. 1702–1706, 2022. DOI: 10.1109/TCSII.2021.3109973.
8. S. Zhang, X. Hu, Z. Liu, et al., “Deep neural network behavioral modeling based on transfer learning for broadband wireless power amplifier,” IEEE Microwave and Wireless Components Letters, vol. 31, no. 7, pp. 917–920, 2021. DOI: 10.1109/LMWC.2021.3078459.

9. A. Fischer-Bühner, L. Anttila, M. Turunen, M. Dev Gomony, and M. Valkama, “Augmented phase-normalized recurrent neural network for rf power amplifier linearization,” IEEE Transactions on Microwave Theory and Techniques, vol. 73, no. 1, pp. 412–422, 2025. DOI: 10.1109/TMTT.2024.3484581.
10. L. Kouhalvandi, O. Ceylan, and S. Ozoguz, “Automated deep neural learning-based optimization for high performance high power amplifier designs,” IEEE Transactions on Circuits and Systems I: Regular Papers, vol. 67, no. 12, pp. 4420–4433, 2020. DOI: 10.1109/TCSI.2020.3008947.
11. C. Belchior, L. C. Nunes, P. M. Cabral, and J. C. Pedro, “Towards the automated rf power amplifier design,” in 2023 19th International Conference on Synthesis, Modeling, Analysis and Simulation Methods and Applications to Circuit Design (SMACD), 2023, pp. 1–4. DOI: 10.1109/SMACD58065.2023.10192148.
12. L. Kouhalvandi and S. D. Guerrieri, “Modeling of hemt devices through neural networks: Headway for future remedies,” in 2023 10th International Conference on Electrical and Electronics Engineering (ICEEE), 2023, pp. 261–267. DOI: 10.1109/ICEEE59925.2023.00054.
13. L. Kouhalvandi and S. D. Guerrieri, “Nonlinear behavioral modeling of fets: Toward the implementation of deep neural networks through large signal data and eda tools,” in 2024 19th European Microwave Integrated Circuits Conference (EuMIC), 2024, pp. 307–310. DOI: 10.23919/EuMIC61603.2024.10732844.
14. P. Chen, J. Xia, B. M. Merrick, and T. J. Brazil, “Multiobjective bayesian optimization for active load modulation in a broadband 20-w gan doherty power amplifier design,” IEEE Transactions on Microwave Theory and Techniques, vol. 65, no. 3, pp. 860–871, 2017. DOI: 10.1109/TMTT.2016.2636146.
15. P. Chen, B. M. Merrick, and T. J. Brazil, “Bayesian optimization for broadband high-efficiency power amplifier designs,” IEEE Transactions on Microwave Theory and Techniques, vol. 63, no. 12, pp. 4263–4272, 2015. DOI: 10.1109/TMTT.2015.2495360.
16. Ampleon, https://www.ampleon.com/, Accessed: 2025-02-23.
17. S. Yarman, “Design of ultra wideband power transfer networks,” New York, NY, USA: Wiley, 2010. DOI: 10.1002/9780470688922.