A Decision-making Module for Fertilization and Irrigation Control System in Rice Farming Using Markov Chain Process and SARSA Algorithms

The merging of modern sensor technologies, the Internet, and advanced control of irrigation and fertilization with an Internet of Things (IoT) approach allow a relatively precise control of agriculture.This IoT approach can thereby increase the resilience of agricultural systems in the face of complex demands for water and fertilizer use, even in countries such as Rwanda with low levels of economic development as long as appropriate and low cost technologies are used.In this work, we add to our previous IoT design for an irrigation system by adding a fertilization system.The proposed low cost system will provide individual farmers fertilization and irrigation options informed by reservoir capacity, water level, predicted rainfall, and temperature along with soil condition and pH.IoT data are assessed in the context of rice growth stages as a Markov Chain process, or in the case of IoT system fault, assessed using the SARSA temporal difference technique.Simulations for the Muvumba Valley rice project in Northeast Rwanda demonstrate the potential of this IoT system to increase rice yield while decreasing fertilizer and water use.