Automated Design of Dielectric Transmitarray Antenna using Deep Learning

This study presents a novel method for the automated design of a transmitarray antenna. By constructing forward and inverse network, the proposed method can rapidly determine the geometric dimensions of the transmitarray antenna. Approximately 2000 groups of datasets were utilized in our work, in which 1800 groups were used for training and 200 groups for testing. Test results demonstrate the model's ability to accurately predict the geometric size of the transmitarray antenna. Four examples with different parameters are shown and their minimal discrepancies between the neural network and the full wave simulation reveal a high level of agreement. The proposed method provides an intelligent and efficient approach to design transmitarray antennas using deep learning.