作者: Ying Wang , Ming Yu , Humayun Kabir , Qi-Jun Zhang
DOI: 10.1109/MWSYM.2006.249539
关键词:
摘要: In this paper, neural network modeling techniques are applied to the design of waveguide dual-mode pseudo-elliptic filter. A hybrid approach is developed where networks and filter coupling matrix combined in an innovative way deliver speed accuracy overall design. Filter structure decomposed into modules representing each mechanism. Generalized scattering matrices (GSM) calculated using mode-matching method. Equivalent circuit parameters, such as value insertion phase lengths then extracted from EM data. Neural models for parameters individual module instead direct GSM. Good agreement obtained between EM-based narrow-bandwidth four-pole Ku band bandpass designed trained NN models, simulated optimized full model (HFSS). The difference dimensions within 0.01? all dimensions, which demonstrates that capable achieving with superior computation speed.