Edge-Based Intelligent Irrigation System Using Iot and AI

Abstract

This thesis presents an intelligent soil moisture monitoring and irrigation management system that integrates the Internet of Things (IoT), Artificial Intelligence (AI), and Edge Computing. The proposed architecture uses a dual-layer AI approach: the first model predicts future soil moisture levels based on historical environmental and sensor data, while the second model makes irrigation decisions by triggering the water pump when predictions fall below a critical threshold. The required water volume is calculated using a tailored formula to meet crop-specific needs. Both AI models are deployed locally on an Edge device (field-based PC), enabling real- time decision-making without reliance on cloud connectivity. This ensures low latency, improves system resilience, and supports autonomous irrigation control. Experimental results demonstrate accurate moisture forecasting and optimized water usage, making the system highly relevant for sustainable precision agriculture, especially in water-scarce regions.