Abstract
<jats:p>Calculating the gain of an antenna array is a critical task in the design and operation of modern radio, radar, and electronic warfare systems. This parameter makes it possible to quantify the efficiency of electromagnetic energy concentration in a given direction, which directly determines the communication range, noise immunity and energy efficiency of the system. In this paper, we propose a method for calculating the AFAR gain using neural networks. The approach is based on training a model on synthesized data, where the input parameters are the lattice geometry, amplitude-phase distribution and scanning angles, and the target variable is the gain value. The trained model demonstrates the ability to instantly predict antenna characteristics for arbitrary configurations. The advantage of the method is a significant reduction in computational costs compared to traditional electrodynamic calculations while maintaining satisfactory accuracy. The developed approach opens up opportunities for creating adaptive directional pattern control systems in real time and effectively optimizing antenna array parameters under changing operating conditions.</jats:p>