Machine Learning-Assisted Optimization of a Metasurface-Based Directly Modulating Antenna

A directly modulating antenna using metasurfaces is optimized using the surrogate model assisted differential evolution for antenna synthesis (SADEA) method and simulated. Metasurface modulation holds promise as an energy efficiency transmitter technology, but suffers from modulation distortion and many differing parameters, making achieving good designs difficult. The algorithm used here, SADEA, obtained a design that shows improvement over conventional design techniques, producing amplitude variation of 1.8 dB over 360° and an average efficiency of 65%, up from 50% obtained by the standard model.