Design of Wideband Electromagnetic Absorbers Based on Deep Learning

This paper presents a deep learning-based multilayer perceptron (MLP) neural network approach for the design and optimization of electromagnetic absorbers operating in the 20-30 GHz frequency range. This method offers significant advantages in terms of miniaturization and wide bandwidth. By training the MLP neural network, the absorber structure is automatically optimized within this frequency band. The predicted results are in strong agreement with full-wave simulation outcomes. The findings demonstrate that this approach effectively enhances absorber efficiency, providing a novel solution for electromagnetic compatibility and stealth technology applications.