| dc.description.abstract | Network technology continues to undergo various changes, especially with the emergence of 5G technology, which is the latest generation with specifications that are significantly better than previous generations. One important component in the implementation of 5G is the antenna, particularly the microstrip antenna, which is known for its high gain and directivity used in various applications, including satellite navigation and telecommunications. This research aims to design and optimize a rectangular microstrip patch antenna that operates at frequencies of 2.6 GHz and 3.4 GHz for 5G technology applications. The optimization process is carried out using a Decision Tree algorithm based on Machine Learning to predict the optimal dimensions of the antenna, which is then simulated using CST Studio Suite software. At a frequency of 2.6 GHz, measurements with the VNA showed a return loss of -15,17 dB, VSWR of 1,30, and a bandwidth of 184 MHz, while CST simulation provided a return loss of -24,408 dB, VSWR of 1,1281, bandwidth of 3,930 MHz, and gain of 3,838 dBi. At a frequency of 3.4 GHz, VNA testing yielded a return loss of -19,8 dB, VSWR of 1,22, and a bandwidth of 75 MHz, whereas CST simulation resulted in a return loss of -31,2561 dB, VSWR of 1,05627, bandwidth of 5,894 MHz, and gain of 3,58 dBi. The predictions from the Machine Learning model showed consistent results with a return loss at 2.6 GHz of -25.6573 dB, VSWR of 1,1157, bandwidth of 4,057 MHz, and gain of 3,984 dBi. At 3.4 GHz, the predicted return loss was -33.326 dB, VSWR of 1.02513, bandwidth of 7,362 MHz, and gain of 3,895 dBi. The accuracy of the Decision Tree model was evaluated using Mean Squared Error (MSE) and Mean Absolute Error (MAE), with an MSE of 6.33 and MAE of 0.68 at 2.6 GHz, and an MSE of 1.45 and MAE of 0.50 at 3.4 GHz. | en_US |
