Abstract
Accurate forecasting of water requirements is crucial for optimizing irrigation and water preservation. However, the Food and Agriculture Organization(FAO Irrigation and Drainage paper 56) Penman-Monteith(PM) model is observed as the highest quality method for evapotranspiration (EVT0 ) forecasting. However, using the PM model is often restricted by the need for predicted climatic factors, particularly solar radiation. This research article presents a real-time intelligent watering system for coriander plants that can be monitored using smartphones. The system uses a hybrid machine-learning technique and Internet of Things (IoT) sensors to sense weather circumstances directly from the crop field. Nine distinct hybrid neural network models ((HML1, HML2 …, HML9)) are developed to predict water requirements using climate and environmental variables. These models are optimized using a genetic algorithm to achieve optimal efficiency. The EVT0 forecasts of the proposed approach are being compared against the standard FAO56 Penman-Monteith technique. An in-depth analysis of the highly successful HML4 model is conducted, and the findings are used in a developed Android application that enables real-time monitoring. In addition, the most favourable parameters are determined to achieve even more improved outcomes. This intelligent irrigation system can significantly minimize flood irrigation, water consumption, and labour expenses.