Design of A Smart Cat Feeding System Using The Sugeno Fuzzy Logic Method Within An IoT Framework

Abstract

The increasing number of cat owners often conflicts with busy schedules, leading to irregular feeding and health risks like obesity or malnutrition. Current automatic feeding systems are generally schedule-based without adjusting portions to individual nutritional needs. This study aims to design an Internet of Things (IoT)-based automatic cat feeder system integrating the Sugeno Fuzzy Logic method to determine feed portions adaptively. The system utilizes a NodeMCU ESP32 as the main controller, three load cells to weigh the cat, food portion, and remaining water, alongside an HC-SR501 PIR sensor detecting daily activity levels. The input variables for fuzzy logic include age, body weight, and activity level, which produce a specific feed portion output (Small 35g, Medium 70g, Large 105g) using an auger drive mechanism. Test results indicate highly precise performance. Weighing accuracy reached 98.62% for cat weight, 98.63% for food portions, and 98.87% for water volume. Furthermore, validating the Sugeno Fuzzy Logic algorithm on the microcontroller against MATLAB simulations yielded a 100% accuracy rate with zero computational error (MSE=0, RMSE=0, R²=1). Feed realization by the actuator achieved 98.85% accuracy. The system proved reliable in maintaining punctuality with an average delay of 11.07 seconds, demonstrating good network resilience through local storage during offline states. This implementation facilitates owners in preventively maintaining the welfare and nutritional balance of their pet cats.