Machine-learning Assisted Synthesis for Series-Fed Microstrip Array

A machine-learning assisted (MLA) synthesis for the series-fed microstrip array method is proposed. The series-fed microstrip array is divided into three types of base element (BE). Each BE is individually sampled and trained, significantly reducing the calculated budget. Once the design parameter of the series-fed microstrip array is given, the S-parameter and radiation pattern of the array can be approximated by the cascade of three types of BEs. Two practical array synthesis examples are given to verify the superiority of the algorithm.