A New Neural Network Model for Accurate Analysis of Microstrip Filters on PBG Structure

The performance of planar circuits, like microstrip filters, has been improved by the use of PBG structures. Usually, the development of fullwave analyses for calculating S parameters of microstrip circuits based on numerical methods (such as FDTD and MoM) is a hard task. In this work, an accurate and efficient neural model is proposed for the analysis of microstrip filters on PBG structures, at the microwave band. The simulation performed showed that the proposed ANN paradigm for microstrip filters provides numerical results that are in excellent agreement with measured results available in the literature. Furthermore, it has been observed that the neural model accomplishes a good generalization in interest regions out of training set, where there are no measured results available.