Design Of Microstrip Antenna Using HFSS Based On Artificial Neural Network

Microstrip antenna is a kind of antenna with the advantages of small shape,lightweight,easy manufacture,low price,compatibility with integrated circuit devices and surface adaptability.Nowadays,users in the world have higher and higher requirements for the microstrip antenna system,which is very important to improve the manufacturing capacity of the microstrip antenna system.Therefore,effective,fast and reliable patch antenna design technology is needed.In this paper,rectangular patch microstrip antennas based on the Artificial Neural Network are designed.In this thesis,a rectangular patch microstrip antenna through High-Frequency Structure Simulator HFSS is designed.The proposed antenna operates at 9.5 GHz frequency which is X-band radar frequency.Furthermore,two excitation techniques are applied to the antenna,which is totally different in geometry,that is microstrip line feeding and coaxial feeding technique.In both techniques,the antenna size is the same.In this work the results of two different feeding techniques were compared.The rectangular patch microstrip antennas for X-band frequency were designed in this thesis,which is from 8 GHz to 12 GHz frequency.The proposed antenna is designed by using HFSS Software based on the method of moment(MOM).Applied artificial neural network(ANN)computing the resonant frequency of rectangular microstrip antenna.Based on the backpropagation(BP)technique,the multilayered perceptron(MLP)model has been used to trained and test the resonant frequency of a rectangular microstrip antenna.Find 85 resonant frequency of rectangular patch microstrip antenna in which 73 data are used for training and 12 data are used for the test,with different dimensions and electrical parameter,applied different learning algorithms for training to achieve the best result.Levenberg Marquardt(LM)algorithm is used because it is more intelligent as compared to the other algorithms.Compare performance with each other in which the best result obtained from Levenberg Marquardt(LM).Thus,the result achieved in this study showed that the ANN model of Levenberg Marquardt(LM)algorithm successfully use the resonant frequency of a rectangular patch microstrip antenna,which give best accuracy.