Intelligent Hydroponic Crop Management System Using Artificial Intelligence, Digital Twin Technology, Fuzzy Logic Control, and Chatbot Assistance

Abstract

This thesis presents the design and implementation of an intelligent hydroponic system combining artificial intelligence and IoT technologies. The first chapter establishes the theoretical background of the project. The second chapter describes two AI models: a plant disease detection model built upon three architectures (Custom CNN, EfficientNet-B3, and ResNet-50), evaluated on 20 disease classes covering tomato, lettuce, and strawberry plants, and a crop yield estimation model based on YOLO. Both models achieved strong performance. The third chapter presents the system’s digital twin, integrating a Mamdani fuzzy logic controller for automatic nutrient and pH regulation, along with an embedded conversational assistant, HydroBot, enabling intuitive system monitoring and interaction.