volume | PIER Journals

Abstract

This study introduces a deep neural network architecture tailored for accurately modeling the parameters of microwave components, with a specific focus on waveguide filters. Unlike simpler neural networks, this architecture is designed to handle the complex relationships that are prevalent in microwave engineering. The model's inputs include the filter's geometric variables and frequency, whereas the S-parameters serve as outputs. To effectively capture these relationships, the Rectified Linear Unit (ReLU) activation function was employed, which is known for its efficiency in managing a significant number of training parameters. This choice allowed the model to better grasp the intricate connection between the S-parameters and geometric variables, and the relationship was found to be more complex than that with frequency. The primary goal is to reduce the overall count of training parameters within the deep neural network while maintaining a level of accuracy similar to that of fully connected neural networks. This study demonstrates the effectiveness of this approach through waveguide filter parametric modeling, highlighting its capacity to accurately model and optimize the electromagnetic response of the filter.

1. Farabet, Clement, Camille Couprie, Laurent Najman, and Yann LeCun, "Learning hierarchical features for scene labeling," IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 35, No. 8, 1915-1929, Aug. 2013.
doi:10.1109/tpami.2012.231 Google Scholar

2. Chen, Dongpeng and Brian Kan-Wing Mak, "Multitask learning of deep neural networks for low-resource speech recognition," IEEE/ACM Transactions on Audio, Speech, and Language Processing, Vol. 23, No. 7, 1172-1183, Jul. 2015.
doi:10.1109/taslp.2015.2422573 Google Scholar

3. Collobert, Ronan and Jason Weston, "A unified architecture for natural language processing: Deep neural networks with multitask learning," Proceedings of the 25th international conference on Machine learning, 160-167, Jul. 2008.
doi:10.1145/1390156.1390177

4. Bengio, Yoshua, "Learning deep architectures for AI," Foundations and Trends in Machine Learning, Vol. 2, No. 1, 1-127, 2009.
doi:10.1561/2200000006 Google Scholar

5. Huang, An-Dong, Zheng Zhong, Wen Wu, and Yong-Xin Guo, "An artificial neural network-based electrothermal model for GaN HEMTs with dynamic trapping effects consideration," IEEE Transactions on Microwave Theory and Techniques, Vol. 64, No. 8, 2519-2528, Aug. 2016.
doi:10.1109/tmtt.2016.2586055 Google Scholar

6. Cho, Kyunghyun, B. Van Merriënboer, C. Gulçehre, D. Bahdanau, F. Bougares, H. Schwenk, and Y. Bengio, "Learning phrase representations using RNN encoder-decoder for statistical machine translation," Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), 1724-1734, 2014.

7. Liu, Wenyuan, Weicong Na, Lin Zhu, Jianguo Ma, and Qi-Jun Zhang, "A Wiener-type dynamic neural network approach to the modeling of nonlinear microwave devices," IEEE Transactions on Microwave Theory and Techniques, Vol. 65, No. 6, 2043-2062, Jun. 2017.
doi:10.1109/tmtt.2017.2657501 Google Scholar

8. Jin, Jing, Chao Zhang, Feng Feng, Weicong Na, Jianguo Ma, and Qi-Jun Zhang, "Deep neural network technique for high-dimensional microwave modeling and applications to parameter extraction of microwave filters," IEEE Transactions on Microwave Theory and Techniques, Vol. 67, No. 10, 4140-4155, Oct. 2019.
doi:10.1109/tmtt.2019.2932738 Google Scholar

9. Zhao, Ping and Ke Wu, "Homotopy optimization of microwave and millimeter-wave filters based on neural network model," IEEE Transactions on Microwave Theory and Techniques, Vol. 68, No. 4, 1390-1400, Apr. 2020.
doi:10.1109/tmtt.2019.2963639 Google Scholar

10. Feng, Feng, Venu-Madhav-Reddy Gongal-Reddy, Chao Zhang, Jianguo Ma, and Qi-Jun Zhang, "Parametric modeling of microwave components using adjoint neural networks and pole-residue transfer functions with EM sensitivity analysis," IEEE Transactions on Microwave Theory and Techniques, Vol. 65, No. 6, 1955-1975, Jun. 2017.
doi:10.1109/tmtt.2017.2650904 Google Scholar

11. Zhang, Q. J. and Kuldip C. Gupta, Neural Networks for RF and Microwave Design, Artech House, 2000.

12. Zhang, Qi-Jun, Kuldip C. Gupta, and V. K. Devabhaktuni, "Artificial neural networks for RF and microwave design-from theory to practice," IEEE Transactions on Microwave Theory and Techniques, Vol. 51, No. 4, 1339-1350, Apr. 2003.
doi:10.1109/tmtt.2003.809179 Google Scholar

13. Karahan, Emir Ali, Zheng Liu, Aggraj Gupta, Zijian Shao, Jonathan Zhou, Uday Khankhoje, and Kaushik Sengupta, "Deep-learning enabled generalized inverse design of multi-port radio-frequency and sub-terahertz passives and integrated circuits," Nature Communications, Vol. 15, No. 1, 10734, 2024.
doi:10.1038/s41467-024-54178-1 Google Scholar

14. Javadi, Sara, Behrooz Rezaee, Sayyid Shahab Nabavi, Michael Ernst Gadringer, and Wolfgang Bösch, "Machine learning-driven approaches for advanced microwave filter design," Electronics, Vol. 14, No. 2, 367, 2025.
doi:10.3390/electronics14020367 Google Scholar

Dr. Nour El Houda Sara Senasli

Dr. Bouhafs Bouras

Dr. Mohammed Chetioui

Dr. Lamia Senasli

Dr. Mehdi Damou