부산대학교 도서관

Farmers in developing nations are faced with limited technology for smart and economical ways of irrigation. This is worsened by the high cost of imports since most of these nations are import-dependent. Increasing agricultural yield to meet the growing population in most of these nations has led their governments to pursue cost-effective ways of managing water and other resources required for optimum agricultural yield. In this study, the Internet of Things (IoT) and edge computing are proposed to provide an intelligent irrigation system to bridge the irrigation gap. This research was carried out using a greenhouse with a controlled environment i.e., the ambient temperature, air humidity, ultraviolet radiations (UV), soil temperature, and soil moisture content were all monitored and measured in real-time within the controlled environment. These data were collected using field sensors processed using IoT -enabled microcontrollers (Arduino kits). The data were processed via Amazon Web Services (A WS) for ontological capabilities of the system consisting of computational capabilities of IoT server and Edge Server. Internet-enabled mobile devices were used to monitor and adopt preferential yield rates for cultivation. Also, an irrigation sequencing algorithm was developed, which uses predictive machine learning K-Mean Clustering Algorithm to predict a more accurate soil moisture content with minimum standard error and a lower mean square error of 0.11 to the predicted soil content compared to when only SVR with mean square error of 0.12.