Deep learning base modified MLP model for precise scattering parameter prediction of capacitive feed antenna
Citation
Çalık, N., Belen, M. A., & Mahouti, P. (2020). Deep learning base modified MLP model for precise scattering parameter prediction of capacitive feed antenna. International Journal of Numerical Modelling: Electronic Networks, Devices and Fields, 33(2), e2682.Abstract
The relations between the antennas' geometrical parameters and design specifi-cations usually consist of linear and nonlinear components. Especially with theincrease of the requested performance measures, the design procedure becomesmuch more complex due to the conflicting performance criteria or designlimitations. To achieve a design with high performance with feasible designparameters, a fast, accurate, and reliable design optimization process is required.Herein, to have a fast, accurate, and high-performance capacitive-feed antennamodel to be used in design optimization problems, a modified multi-layerperceptron (M2LP) model has been proposed. The M2LP is an equivalent con-volutional neural network (CNN) model of a standard multilayer perceptron(MLP), where instead of traditional training parameters of MLP, more advancedtraining parameters of CNN models such as batch-norm layer, leaky-rectifiedlinear unit (ReLU) layer, and Adam training algorithm had been used. Further-more, the M2LP model had been used in a design optimization process and theobtained optimal antenna had been prototyped using 3D printing technology forjustification of the proposed M2LP model with experimental results. As can beseen from the results, the proposed M2LP model is a fast, accurate, and reliableregression model for design optimization of microwave antennas.