A Low Cost Smart Irrigation Planning Based on Machine Learning and Internet of Things

Abstract

In an agriculture system major issues of irrigation systems for plants water supply is a critical factor. A significant amount of freshwater is required for this task but after the utilization of water in the irrigation process it is being polluted. In addition, the excessive use of water during the irrigation process can negatively affect crop production. Therefore, we need to provide a balanced amount of water for effective crop production and conservation of water. In this paper, we proposed low-cost irrigation planning with two key aims: first is to reduce the installation and maintenance costs of data collection in innovative irrigation systems and second is to control the valve for water supply automatically. In this context, we first provide a review of recent irrigation systems based on the Internet of Things (IoT) and Machine Learning (ML). Next, we introduce a working plan to collect crop water requirements using a soil moisture sensor. Then, an algorithm is proposed to decide the water supply for water treatment. Finally, the experiments are conducted on the samples collected from the farmland of wheat crops. Additionally, two different scenarios are considered to collect the water requirement samples. Based on the experimental and theoretical analysis of water requirements the proposed irrigation system can reduce the water demand by up to 25% as compared to traditional ways of irrigation. Moreover, in comparison of popular valve automation system the proposed multiple valve based system reduces the amount of water wastage up-to 22%. Therefore by utilizing the advance computational techniques (IoT and ML), we can reduce the cost of irrigation system and planning.