An Effective Artificial Neural Network-Based Method for Linear Array Beampattern Synthesis

The beampattern synthesis of antenna array is always known as a computationally cost task that needs efficient strategies to deal with. In this article, a novel artificial neural network (ANN)-based array synthesis method is proposed. By establishing an encoder-decoder-based ANN framework, the mask-constrained beampattern in terms of focused or shaped beam for linear arrays with arbitrary given array geometry is successfully synthesized. More in detail, the encoder serves as an array synthesizer, while the decoder behaves as an array analyzer. Thanks to the designed pretrained decoder, such an approach is computationally efficient so that real-time array synthesis can be potentially achieved. Moreover, the proposed method allows one to consider both ideal and actual linear arrays with mutual coupling effects and nonidealities. The results of a wide numerical validation involving amplitude-only, phase-only, and amplitude-phase excitation syntheses are presented to assess the flexibility and the versatility of the proposed method also in comparison to the competitive state-of-the-art synthesis techniques.