IoT-based Precision Agriculture System using LoRa with Enhanced Soil Fertility Prediction and Automated Irrigation

Agriculture serves as the backbone of Indian Economy, with more than 50% of the population involved in farming activities. Precision agriculture, driven by advancements in IoT and machine learning, presents groundbreaking opportunities to enhance agricultural productivity and sustainability. This paper introduces an innovative system for soil fertility prediction and irrigation management, combining IoT-based real-time data with predictive analytics. Using real-time data acquisition capabilities, the LoRa-based IoT network system collects vital parameters such as the pH of the soil, its moisture content, the temperature, and even humidity levels. The central component of the system for soil fertility prediction leverages a stacked ensemble machine learning model with a high accuracy of 93.37%, along with techniques such as SMOTE to tackle class imbalance present in the dataset and polynomial-based feature engineering to further boost model performance. An optimized Long Short-Term Memory model is employed for irrigation management, providing accurate irrigation requirements with exceptional accuracy, reaching 99.53%, integrating both current sensor data and weather forecasts. This novel approach marks a major move toward sustainable and efficient farming practices in an increasingly digital world.