Effective modeling of filters in the design and optimization of multiplexers

In this paper, an overview of our recent development in the effective modeling of microwave filters is presented, highlighting the application of neural network (NN) modeling technique in the design and optimization of multiplexers. For this application, NN models are based on much simplified EM data of only critical filter structures. The cost of model development is therefore significantly reduced. Furthermore, combination of EM, NN and circuit models enables effective design and optimization of complex circuits. The efficiency of the modeling techniques has been demonstrated through a number of medium to large size multiplexers with dual-mode circular waveguide channel filters.