Machine Learning Based Approach to Accurately Model and Optimize Microwave Filters

Lately, Artificial Neural Networks (ANNs) have been heavily used for radio frequency and microwave circuits and devices modelling. In this context, ANN based models are developed and assessed in this paper for the modelling and optimization of microwave filters. Four filter designs namely Low-Pass Filter (LPF), High-Pass Filter (HPF), Band-Pass Filter (BPF) and Band-Stop Filter (BSF) are developed and fabricated. The obtained S-parameters behaviors for a broad 0–10 GHz frequency spectrum are analyzed and simulated using ANNs. The developed models demonstrate excellent simulation accuracy, and regression score (R2) more than 99 % for all the filter designs,