Optimized Evolutionary Gravitational Neocognitron Neural Network for frequency reconfigurable antenna design

Frequency reconfigurable antenna design using Evolutionary Gravitational Neocognitron Neural Network optimized with Balancing Composite Motion Optimization (FRA-EGNN-BCMO) proposed here. Frequency reconfigurable antenna works at various frequency band at suitable resonance frequency based on choice over switching circuit. For designing the Frequency reconfigurable antenna, the maximum and minimum are needed to optimize. The antenna impedance and magnitude are depending on and of antenna. Here, and are considered as the impedance matching parameters. Hence, Evolutionary Gravitational Neocognitron Neural Network (EGNN) used for optimizing and impedance matching parameters. In general, EGNN do not show any adoption of optimizing systems to compute optimum parameter to make sure exact design. Therefore, in this work, the proposed Balancing Composite Motion Optimization (BCMO) used to optimize weight parameter of EGNN. The proposed FRA-EGNN-BCMO method performance system is analyzed under several metrics like, Bandwidth (MHz), Reflection Coefficient (in dB), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), forecast accuracy. Simulation results of proposed FRA-EGNN-BCMO design provide better predicted accuracy 34.79%, 24.88% and 33.09% compared with existing methods, like FRAD-BPNN, FRAD-PSADEA and FRAD-ITLBO respectively.