UDC 621.396.6
ALGORITHM FOR NEURAL NETWORK MODELING OF COUPLED-LINE FILTERS
T. D. Luu, post-graduate student, RSREU, Ryazan, Russia;
orcid.org/0000-0003-3347-1469, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
E. P. Vasiliev, Doctor in Technical Sciences, Professor, Department of Space technologies, RSREU, Ryazan, Russia;
orcid.org/0000-0003-2747-7012, e-mail: This email address is being protected from spambots. You need JavaScript enabled to view it.
This paper discusses an algorithm for modeling coupled-stripline filters (CSL) using artificial neural networks (ANN). The approaches to modeling filters using ANNs are analyzed, focusing on various filter designs and frequency ranges. The aim of this paper is to develop an ANN-based algorithm for modeling and optimizing CSL filters, which significantly reduces design time compared to traditional electrodynamic simulators while maintaining high calculation accuracy. Input parameters of the model are filter dimensions and electrophysical parameters of substrate, and output parameters are transmission coefficients S21 and reflection coefficients S11. Training database for a neural network model is generated using electrodynamic simulator. A 5-resonator bandpass filter (BPF) based on a coupled-line transistor with center frequency of 16 GHz and a –3 dB bandwidth of 32 % is con sidered. The developed neural network model demonstrates high accuracy in matching calculation results with electrodynamic simulators, while significantly reducing computational costs.
Key words: simulation, coupled-line filter, artificial neural network, S-parameters.