Internet of Thing (IoT) Enabled Smart Irrigation System

Description

Food and water are two essential components of human life. No one can sustain with water and food. Therefore, for human sustainability and to fulfill the increasing food and water demand it is essential to improve the food production quality and quantity. In addition, it is also required to preserve the water. However, by using new sensor technology and Machine Learning (ML) techniques smart farming and precision farming techniques are developed. Both technologies assist farmers in different phases and activities in farming to improve crop productivity and preserve resources. However, the deployment and maintenance of both technologies are much expensive therefore in this project a smart irrigation system has been proposed to reduce the cost of deployment and maintenance.

The proposed smart irrigation system utilizes the concept of the Internet of Things (IoT) and Machine Learning. The IoT infrastructure is used to sense, collect, and communicate the farm conditions in terms of water requirements. In this experiment, the temperature and moisture dataset has been considered for demonstrating the data collection from the farm. The IoT sensor-based data has a time series forecasting problem therefore time series decomposition has been performed and water requirement trends have been studied. Next, time series data has been transformed to learn the water requirements of the plants and predict the water treatment plan.

In this context, an Artificial Neural Network architecture has been implemented to train and predict the water treatment plan. Next, an algorithm is developed to sense, process, and predict the water treatment plant without any human intervention. Finally, the performance of the model has been measured in terms of precision, recall, f-score, and accuracy. Based on the overall efforts the model is found acceptable and applicable to real-world use. The model helps to preserve human resources, water resources, and expensive time. Finally, the conclusion and future work is discussed.