Optimasi Antena Mikrostrip Patch Rectangular pada Frekuensi 1646 MHZ untuk Aplikasi Emergency Position Indicating Radio Beacon (EPIRB) Berbasis Machine Learning

Abstract

In the era of advanced technology, communication has become an important aspect of everyday life, especially in emergencies. The Emergency Position-Indicating Radio Beacon (EPIRB) is an important application used to provide location and signal when an emergency situation occurs at sea. EPIRB performance depends heavily on its size and configuration. Modern communication technology has led to the need for multifunctional and integrated antennas. The Microstrip antenna has a high advantage and directivity used in a variety of applications such as satellite navigation and telecommunications. Machine learning algorithms are considered effective in determining the optimal value of the size of the rectangular patch microstrip antenna for EPIRB applications. Based on comparison of parameters between simulation, machine learning, and testing with VNA Master obtained return loss -31,473528 dB, VSWR 1,0548405, bandwidth 62,1 MHz, and gain 2,484 dBi on CST Studio Suite simulation; return loss -31,736658 dB, VSWR 1,054731, bandwidth 61,6 MHz, and gain 2,485 dBi on machine learning with Extreme Gradient Boosting algorithm (XGBoost); and return loss -12.58 dB, VSWR 1.61, and 51 MHz bandwidth on performance tests with the VNA Master.