Next Generation Antenna Design Synthesis Framework Using k-Nearest Neighbours Algorithm

The state-of-the-art antenna design synthesis and optimization techniques cannot reduce the requirement of high computational resources and domain expertise. Therefore, we suggest a new framework using a machine learning model (k-nearest neighbors algorithm) with the ability to design an antenna with user-defined requirements. To demonstrate the workings of the proposed framework, two specific coplanar waveguide-fed patch antenna designs were developed based on user-defined requirements with almost 100% accuracy and a run time less than 5 seconds.