<![CDATA[The management of feed is a pivotal element in the success of catfish farming. However, in practice, the management of feed is predominantly conducted manually, which can result in feed wastage and inconsistency in feeding times. The objective of this study is to develop and evaluate a prototype of a real-time fish feed control system based on the Internet of Things (IoT) called Feedez. This system is designed to support the automation of feeding in catfish farming. The method employed is evolutionary prototyping, which encompasses a systematic sequence of steps including needs analysis, architecture design, prototype implementation, functional testing, and evaluation and iteration of improvements. The primary contribution of this research is the development of a system that exhibits high operational resilience through a dual data persistence mechanism. In addition to synchronization on the database server, the feeding schedule configuration is stored locally in the Arduino EEPROM memory to ensure security even in the event of network failure or power outage. Additionally, Feedez is equipped with multi-level (5-level) feed discharge speed control using pulse-width modulation (PWM) signals to ensure even feed distribution in large ponds. The experimental results indicate that the system possesses the capacity to execute feed scheduling in a timely manner, with an average output of 110 grams per minute for a feed with a diameter of 3 millimeters. The system's capacity to function autonomously in the event of an internet connection loss is ensured by the synchronization of the RTC module and local memory. This ensures the production of an even distribution of feed in a 4-meter diameter biofloc pond. These findings demonstrate that the Feedez prototype exhibits excellent operational reliability as an effective feed automation solution, particularly within the context of intensive catfish farming.]]>
