A Smart Water Management in a Paddy Field Using IOT Technology and Machine Learning

Abstract

Alternative Wetting and Drying (AWD) is one of favorable rice growing methods since it lowers the costs for about 40% from less consumption of water, insecticide and fertilizer. This paper proposes a water management tool to help AWD rice farming. We design the tool with a combination of precision agriculture and smart farm from applying IoT and AI technology. Soil humidity and water level detection sensor are deployed to realize real-time and accurate essential environmental factors in a field while a pipe water-pump operated using electricity is designed to handle water in a field. The pump and sensors are connected with a network for an automatic operation using IoT. For decision-making, supervised learning technique is used to generate a model capable to consider when to operate a water pump from past data. This allows flexibility in generating decision-making rules for variation of locations and soil types with a requirement of having past data. From evaluations, experts satisfied with the proposed tool as the tool is well-designed for safety and proper performance. Regarding usefulness of the tool, experts and end-users rated high score for the ease of usage and its ability to reduce wasting of resource consumption comparing to the traditional method.