Synthesis of Planar Isophoric Non-Uniformly Spaced Array for Microwave Power Transmission via Deep Neural Network

This work presents a synthesis method of planar isophoric sparse array with maximum beam collection efficiency based on a deep neural network (DNN). Due to the powerful fitting capability of neural networks, a DNN framework is established to determine element positions. Crucial constraints including the aperture size, the minimum element spacing, and the maximum sidelobe level outside the receiving region are considered. This method largely reduces the cost and complexity of feeding networks. Numerical results are presented and compared to the state of the art to demonstrate effectiveness and efficiency.