DESIGN AND IMPLEMENTATION OF MECHANICAL, INTERFACE, AND COMMUNICATION FOR SMARTFISHSENSE: A DEEP-LEARNING BASED FISH APPETITE DETECTION SYSTEM

• Main Author: Putra Masyiakh, Oktavian
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/73877
• Institution: Institut Teknologi Bandung

The current global trend in fish consumption is experiencing a steady increase and is projected to continue rising in the coming years. The OECD predicts that worldwide fish consumption will increase by 11% by 2030. Global fish production has also been steadily growing. In Indonesia, national fish consumption increased by 1.48% in 2021. However, domestic fish production has fluctuated and experienced a decline of 2.9% in 2020. Water quality, fish nutrition, and other environmental factors can influence fish production fluctuations. Meanwhile, fulfilling fish nutrition is a significant focus in fish cultivation, but approximately 60% of the provided feed needs improvement, leading to potential issues. This can harm the quality and quantity of harvests and even result in mass fish mortality. Therefore, a system is needed to reduce wastage in fish feeding during cultivation. The authors and the team propose a SmartFishSense system based on the above explanations. SmartFishSense is a fish appetite detection system as an IoT sensor. The TA222301021 final project group develops this system as part of completing the Electrical Engineering Final Project at ITB. In this document, the authors explain the implementation of a user interface for the SmartFishSense product. The interface implementation for SmartFishSense is necessary to enable users to monitor fish appetite without being present at the cultivation site (Remote Monitoring). The interface is designed using the Python programming language, utilizing the Streamlit library to facilitate interface development and the MongoDB database to store and process the acquired data. The SmartFishSense interface is designed as a web development, accessible through the user's internet-connected device website. The authors ensured the interface was user-friendly during development and offered adequate functionality. Consequently, users can monitor and control fish-feeding activities in real time, even when not physically present at the cultivation site. The interface development process involves understanding user needs. We conducted surveys to gain insight into fish cultivation needs and joint problems. By considering input from potential users, we can design an interface that meets practical requirements and provides appropriate solutions. iv In addition to the user interface, the authors implemented a mechanical design as a container casing for SmartFishSense. This container is designed to function as a protective housing for the SmartFishSense device, shielding it from environmental influences that could damage its components. The authors considered material durability against weather, water protection, and security against other external disturbances. In the development of mechanical design, the authors used the Tinkercad 3D design software to create an accurate and realistic model. This model was then 3D printed to ensure the accuracy and quality of the resulting casing. Throughout this process, the authors tested and validated the casing's performance to provide optimal protection for the SmartFishSense device. Through this document, the authors hope to contribute to developing IoT technology for fish cultivation applications. SmartFishSense offers a practical and efficient solution for monitoring fish appetite, a critical factor for successful cultivation. The results of this research are expected to significantly contribute to technology development in fish cultivation and enhance overall fish production efficiency, particularly in Indonesia.