Recent advances in automated multilayer ANN model generation for microwave components

Automated model generation (AMG) is recognized as an efficient method to automatically develop artificial neural network (ANN) models for microwave components. It replaces the complicated manual operations during ANN modeling with a completely automated ANN development framework, thus effectively improving the ANN modeling efficiency. In this paper, a recent advanced AMG method combining interpolation techniques and multilayer neural model structure adaptation is presented and a microwave cavity filter modeling example is used to demonstrate the effectiveness of presented method.