VOI - Voice of Indonesia | RRI - Radio Republik Indonesia

University of Indonesia (UI) students who are members of the UI Autonomous Marine Vehicle (AMV) Team again contributed to achievements in the international arena. A team of 26 students from various cross-faculty study programs at UI designed a ship called ‘MAKARA MH4 XX’. The ship won 3rd place in the Video Skills Category: Sensor Optimization at the 14th International RoboBoat Competition 2021. This victory is a comeback from the AMV UI team after a year of absence from the annual competition. RoboBoat 2021 was participated by 14 teams from 6 countries with 303 participants. They come from high schools and universities at the undergraduate level, and postgraduate level. The competition was held online from March 29 to June 27, 2021. RoboBoat aims to provide a simplified version of the challenges faced in the modern world maritime industry. RoboNation is an institution that focuses on the development of robotic technology based in Florida, United States.

The MAKARA MH4 XX ship is designed by using the Hydrofoil Supported Catamaran (HYSUCAT) concept. The basic principle of HYSUCAT is to utilize a high performance hydrofoil system in the tunnel between the two hulls. The hydrofoil system consists of a main foil placed between and parallel to the bottom of the hull, slightly towards the ship's center of gravity. The team under the guidance of Dr. Muhammad Arif Budiyanto, Head of the Marine Engineering Study Program, Faculty of Engineering, University of Indonesia, chose the HYSUCAT design concept because of several advantages, including that HYSUCAT is much more comprehensive than a catamaran or monohull of comparable size, especially during heavy loads and at high speeds. The performance of the Autonomous Surface Vehicle (ASV), or autonomous surface ship, is critical for competition, especially on speed gate mission tasks. The Skills Video category participated in by MAKARA MH4 XX is intended to highlight how autonomous ships have various input sensors, and can integrate all of these sensors so that the ship carries out its mission quickly and precisely. This integration is expressed in the form of video skills that are collected and then published on the RoboBoat YouTube channel.

Angelita Cindi Viani, AMV UI Team manager explained that the team implemented a new machine learning method, namely transfer learning for the automated detection program from the ship. Transfer learning is the process of building a custom machine learning model, based on a model that has been pre-trained to study thousands of images. According to Angelita, the trained model used is MobileNet, because it has fewer parameters, smaller network size, and faster performance. By using this method, object detection in each mission will be more accurate; the robot will move in a more directed way with a better detection rate. In addition, transfer learning reduces the computational requirements and saves a lot of time and team resources. RoboBoat gives participants the opportunity to design, build, and test an Autonomous Surface Vehicle (ASV), as well as compare the unique systems of each vessel with contestants from around the world.