Machine Learning based soil moisture prediction for Internet of Things based Smart Irrigation System

Internet of things (IoT) and machine learning (ML) based solution are revolutionizing many fields of humankind like transportation, environment, business and agriculture. The fresh water resources, which are already stressed, are being used extravagantly in many countries. The Internet of Things and machine Learning techniques can be used to optimize the water usage in irrigation. This paper presents the application of ML techniques to optimize the irrigation water usage by predicting the future soil moisture of a field in an IoT driven smart irrigation framework. The field data collected from the deployed sensors (air temperature, air humidity, soil moisture, soil temperature, radiation) and the weather forecast data from the Internet are used for predicting the future soil moisture. Multiple ML techniques are analyzed for predicting future soil moisture and the results obtained using GBRT are quiet encouraging. The proposed techniques could be a crucial research front for optimizing the water usage in irrigation.