Gain Bandwidth Enhancement and Sidelobe Level Stabilization of mm-Wave Lens Antennas Using AI-driven Optimization

This paper explores the transformative potential of artificial intelligence (AI) techniques in optimizing the phase distributions of a lens antenna to significantly enhance the gain bandwidth and stabilize the sidelobe levels at the millimeter-wave band. Through an AI-driven antenna design method (self-adaptive Bayesian neural network surrogate-model-assisted differential evolution for antenna optimization (SB-SADEA), specifically), this work obtains a phase distribution that provides a wide gain bandwidth and stable sidelobe levels from 24 to 33 GHz. A lens antenna with 20 × 20 unit cells is implemented based on the phase distribution. Results show a 1-dB bandwidth of 28.2% and the sidelobe levels have also been lowered compared to the reference design. The optimized lens antenna shows a stable gain with a range of 20.13 dB to 22.16 dB from 24 to 33 GHz, in comparison to the reference design that has a gain range of 16.70 dB to 26.43 dB over the same frequency spectrum. The measured results align well with the simulated results, verifying the effectiveness of the AI-driven antenna design optimization technique in enhancing the performance of a lens antenna.