مدل سازی رفتار گذرای وابسته به زمان مدار حلقه قفل فاز دیجیتالی به کمک شبکه¬ی عصبی واحد بازگشتی گیتی
محورهای موضوعی : مهندسی برق و کامپیوتر
سیده فاطمه موسوی قوام آبادی
1
,
سیدعلیرضا صدرالسادات
2
,
علی مفتخرزاده
3
*
1 - دانشکده مهندسی برق، دانشگاه یزد، یزد، ایران
2 - دانشکده مهندسی کامپیوتر، دانشگاه یزد، یزد، ایران
3 - دانشکده مهندسی برق، دانشگاه یزد، یزد، ایران
کلید واژه: مدار DPLL, مدل¬سازی رفتار دینامیکی مدارات غیر¬خطی, محو گرادیان, شبکههای عصبی GRU.,
چکیده مقاله :
امروزه شبکههای عصبی یک ابزار قدرتمند برای مدلسازی مدارهای پیچیده و غیرخطی میباشند. در این مقاله، مدلسازی رفتار گذرای وابسته به زمان مدار حلقه قفل فاز دیجیتالی به کمک دو مدل شبکهی عصبی که وابسته به زمان هستند، انجام شده است. مدلسازی با استفاده از شبکه عصبی RNN با چالش محو گرادیان در مرحله آموزش مواجه میباشد و از نظر سرعت و دقت در مدلسازی عملکرد مناسبی ندارد. برای دستیابی به مقادیر مطلوب خطای آموزش و آزمون، از شبکه عصبی GRU برای مدلسازی استفاده میشود. این شبکه به دلیل وجود دو گیت بهروزرسانی و بازنشانی، قادر میباشد مشکل محو شدن گرادیان را برطرف کرده و مدلسازی قابل قبولی ارائه دهد. مقایسه نتایج این دو شبکه نشان میدهد که روش مبتنی بر ساختار شبکه عصبی گیتی، توانایی و قابلیت بهتری در مدلسازی رفتار مدارهای غیرخطی دارد. در پایان، جهت ارزیابی منصفانه عملکرد مدلها، مقایسهای نیز با ساختار LSTM انجام شده که نتایج آن نشان میدهد شبکه GRU از نظر دقت مدلسازی و سرعت آموزش عملکرد بهتری نسبت به LSTM نیز دارد.
Nowadays, neural networks are a powerful tool for modeling complex and nonlinear circuits. In this paper, the transient time-dependent behavior of a digital phase-locked loop circuit is modeled using two time-dependent neural network models. Modeling with an RNN (Recurrent Neural Network) faces the challenge of gradient vanishing during the training phase and does not perform well in terms of speed and accuracy. To achieve desirable training and testing error values, the GRU (Gated Recurrent Unit) neural network is used for modeling. This network, due to the presence of update and reset gates, can overcome the gradient vanishing problem and provide acceptable modeling results. A comparison of the results of these two networks shows that the method based on the gated neural network structure has superior capability and performance in modeling the behavior of nonlinear circuits.
[1] V. K. Devabhaktuni, M. C. Yagoub, Y. Fang, J. Xu, and Q. J. Zhang, "Neural networks for microwave modeling: Model development issues and nonlinear modeling techniques," International Journal of RF and Microwave Computer-Aided Engineering, vol. 11, no. 1, pp. 4-21, 2001.
[2] Q. -J. Zhang and K. C. Gupta, Neural Networks for RF and Microwave Design (Book+ Neuromodeler Disk). Artech House, Inc., 2000.
[3] W. Liu, W. Na, L. Zhu, and Q.-J. Zhang, "A review of neural network based techniques for nonlinear microwave device modeling," in Proc. IEEE MTT-S Int. Conf. on Numerical Electromagnetic and Multiphysics Modeling and Optimization, 2 pp., Beijing, China, 27-29 Jul. 2016.
[4] A. Rawat, R. Yadav, and S. Shrivastava, "Neural network applications in smart antenna arrays: A review," AEU-International Journal of Electronics and Communications, vol. 66, no. 11, pp. 903-912, Nov. 2012.
[5] S. A. Sadrossadat, P. Gunupudi, and Q. -J. Zhang, "Nonlinear electronic/photonic component modeling using adjoint state-space dynamic neural network technique," IEEE Trans. on Components, Packaging and Manufacturing Technology, vol. 5, no. 11, pp. 1679-1693, Nov. 2015.
[6] W. Na, F. Feng, C. Zhang, and Q. -J. Zhang, "A unified automated parametric modeling algorithm using knowledge-based neural network and l1 optimization," IEEE Trans. on Microwave Theory and Techniques, vol. 65, no. 3, pp. 729-745, Mar. 2016.
[7] H. Zhang, Y. Jing, and P. Zhou, "Machine learning-based device modeling and performance optimization for FinFETs," IEEE Trans. on Circuits and Systems II: Express Briefs, vol. 70, no. 4, pp. 1589-1585, Apr. 2022.
[8] H. Kabir, et al., "Smart modeling of microwave devices," IEEE Microwave Magazine, vol. 11, no. 3, pp. 105-118, 2010.
[9] B. Mutnury, et al, "Macromodeling of nonlinear transistor-level receiver circuits," IEEE Trans. on Advanced Packaging, vol. 29, no. 1, pp. 55-66, Feb. 2006.
[10] M. Noohi, A. Mirvakili, and S. A. Sadrossadat, "Modeling and implementation of nonlinear boost converter using local feedback deep recurrent neural network for voltage balancing in energy harvesting applications," International Journal of Circuit Theory and Applications, vol. 49, no. 12, pp. 4231-4247, Dec. 2021.
[11] R. Dinesh and R. Marimuthu, "An analysis of ADPLL applications in various fields," Indonesian Journal of Electrical Engineering and Computer Science, vol. 18, no. 2, pp. 856-866, May 2020.
[12] P. K. Hanumolu, G. -Y. Wei, U. -K. Moon, and K. Mayaram, "Digitally-enhanced phase-locking circuits," in Proc. IEEE Custom Integrated Circuits Conf., pp. 361-368, San Jose, CA, USA, 16-19 Sept. 2007.
[13] M. Moradi, S. A. Sadrossadat, and V. Derhami, "Long short-term memory neural networks for modeling nonlinear electronic components," IEEE Trans. on Components, Packaging and Manufacturing Technology, vol. 11, no. 5, pp. 840-847, May 2021.
[14] M. Azodi, S. A. Sadrossadat, and Y. Savaria, "Nonlinear circuit macromodeling using new heterogeneous-layered deep clockwork recurrent neural network," IEEE Access, pp. 89506-89519, 2024.
[15] T. Kandadi and G. Shankarlingam, Drawbacks of LSTM Algorithm: A Case Study, Available at SSRN, no. 5080605, Jan. 2025, http://dx.doi.org/10.2139/ssrn.5080605.
[16] H. Ibrahim et al., "Radio frequency energy harvesting technologies: A comprehensive review on designing, methodologies, and potential applications," Sensors, vol. 22, no. 11, Article ID: 4144, 2022.
[17] L. G. B. Rolim, D. R. da Costa, and M. Aredes, "Analysis and software implementation of a robust synchronizing PLL circuit based on the pq theory," IEEE Trans. on Industrial Electronics, vol. 53, no. 6, pp. 1919-1926, Dec. 2006.
[18] S. AI-Araji, Z. M. Hussain, and M. AI-Qutairi, Digital Phase Locked Loops, Architecture and Applications, Springer, Berlin, 2006.
[19] R. J. Baker, CMOS: Circuit Design, Layout, and Simulation, John Wiley & Sons, 2019.
[20] D. Banerjee, PLL Performance, Simulation and Design, Dog Ear Publishing, 2006.
[21] R. I. Abdo, A. A. Khalaf, and H. F. A. Hamed, "A power efficient and fast locking CMOS design of all-digital phase-locked loop," Journal of Advanced Engineering Trends, vol. 42, no. 1, pp. 197-203, Jan. 2022.
[22] Y. Fang, M. C. Yagoub, F. Wang, and Q. -J. Zhang, "A new macromodeling approach for nonlinear microwave circuits based on recurrent neural networks," IEEE Trans. on Microwave Theory and Techniques, vol. 48, no. 12, pp. 2335-2344, Dec. 2000.
[23] Z. Naghibi, S. A. Sadrossadat, and S. Safari, "Time-domain modeling of nonlinear circuits using deep recurrent neural network technique," AEU-International Journal of Electronics and Communications, vol. 100, pp. 66-74, Feb. 2019.
[24] J. Chung, Empirical Evaluation of Gated Recurrent Neural Networks on Sequence Modeling, arXiv preprint arXiv:1412.3555, 2014.
[25] R. Fu, Z. Zhang, and L. Li, "Using LSTM and GRU neural network methods for traffic flow prediction," in Proc. 31st Youth Academic Annual Conf. of Chinese Association of Automation, pp. 324-328, Wuhan, China, 11-13 Nov. 2016.
[26] J. Yuan and Y. Tian, "An intelligent fault diagnosis method using GRU neural network towards sequential data in dynamic processes," Processes, vol. 7, no. 3, Article ID: 152, 2019.
[27] T. Robert, M. Safaryan, I. V. Modoranu, and D. Alistarh, LDAdam: Adaptive Optimization from Low-Dimensional Gradient Statistics, arXiv preprint arXiv:2410.16103, Oct. 21, 2024.