IoT-Enabled Adaptive Watering System With ARIMA-Based Soil Moisture Prediction for Smart Agriculture

Agriculture is the backbone of every country as it fulfills the necessities of human beings as well as animals and also supports the country’s economy in terms of revenue by exporting food products to the international market. But due to rapid growth of industries and population food and water shortage is becoming a very critical issue. Especially in Pakistan shortage of water and unpredicted weather conditions are affecting agricultural areas, and as a result, food shortage is increasing. A solution to handle the water shortage is a controlled watering system for crops. In this paper, we describe two different IoT-based systems for crop soil moisture, soil temperature, and outdoor temperature and humidity monitoring, one is Wi-Fi based while the other is a GSM-based system. This helps us monitor the soil moisture content and other parameters in real-time which will help agriculture researchers and farmers to use water according to system reports. The system uses an ARIMA (autoregressive integrated moving average) machine learning model to predict the soil moisture for the next ten days which will help the farmers to arrange the water in Hyetal areas where canal water is not reachable. The predictions are compared with real-time data and the success rate of the ARIMA ML model is from $80-90\%$ . The proposed research will open the door for agriculture researchers to gather new datasets and predictions by adopting this proposed research which will be beneficial to farmers to cultivate the crops and fruits in hyetal lands according to data recommendation.