Skip to Main Content
This work describes a new class of modular network that has been used to analyze microwave waveguides on electromagnetic bandgap (EBG) structures. The proposed modular configuration employs three or more SF-ANN (sample function artificial neural network) trained by the resilient backpropagation (RProp) algorithm. Two types of waveguides are modeled by this new modular neurotechnique that are the rectangular waveguide and the parallel plate waveguide with EBG structures. The rectangular waveguide uses the UC-EBG structure, which consists of a metal pad and four connecting branches. This device is employed in the development of high-performance and compact circuit components for microwave and millimeter-wave frequencies. The parallel plate waveguide is formed by the microstrip ground plane and a metal plate with EBG elements on top that works like one metal shield. This structure has the capability of suppressing the propagation of unwanted parallel plate modes. Simulations through the proposed neural network models for microwave waveguides on EBG structures gave results in excellent agreement with measured results available in the literature. Furthermore, this technique is able to generalize for new values of input parameters and frequency ranges, validating it as a good estimator into the mapping regions where no previous numerical or experimental information is available.