Back to Publications

Artificial Neural Network Model to Predict the Design Parameters of Inset-fed Microstrip Patch Antenna

Artificial Neural Network Model to Predict the Design Parameters of Inset-fed Microstrip Patch Antenna

Publication Information

Journal/Conference:
8th International Conference on Signal Processing and Integrated Networks (SPIN)
Pages:
943-948

Abstract & Details

This paper demonstrates the use of Artificial Neural Network model to predict the physical parameters of the inset-fed microstrip patch antenna (IMPA) of rectangular shape. The model is proposed with the input variables of frequency, substrate height, and dielectric constant of the substrate whereas predicted output parameters are: patch width, patch length, inset depth, feed line width, and notch gap. Neural Network has been demonstrated using the feed-forward back-propagation algorithm and has been optimized using the “Adam” optimizer. The model is developed by varying the batch size, the number of hidden layers, and the epochs. The developed optimized model showed mean squared error (MSE) of 0.2249, root mean square error (RMSE) of 0.4742, and mean absolute error (MAE) of 0.2075 for the test dataset.