Developing a Mobile-based Educational Game to Enhance Dietary Habits for Type 2 Diabetes Patients through Artificial Intelligence Algorithms

Abstract

This study aimed to design a mobile digital game to assist individuals with type 2 diabetes in better understanding food calories and the glycemic index, ultimately enhancing their management of the condition. The game was developed using fuzzy logic and initially tested in MATLAB 2018 before being converted to C# in Visual Studio and implemented in Unity. To develop the game, food calorie and glycemic index values were integrated into a fuzzy input system, utilizing a triangular membership function. The fuzzy output was translated into a numerical value through the centroid defuzzifier, employing the Mamdani fuzzy inference engine for determination. The outcome of this study was the development of a mobile game named "Diabetic Amoo," specifically designed for diabetic patients. Players advance through the first episode by correctly selecting appropriate food items, while the second stage focuses on educating them about low-sugar and low-calorie foods. Players receive ratings for their choices that range from "very bad" to "very good," with the goal of achieving a "very good" rating. By emphasizing patient education, such games can enhance motivation for self-care and improve adherence to diabetes diets and other health conditions.