A low-cost constant-moisture automatic irrigation system using dynamic irrigation interval adjustment

Abstract

This paper developed a Soil Moisture Forecasting (SMF) model and a Constant-moisture Automatic Irrigation System (CAIS). The SMF model used the soil moisture data at different depths in an experimental plot inside a greenhouse to infer the soil moisture level after a specific interval. CAIS integrated the SMF data with dynamic watering interval adaption to maintain soil moisture at a constant level. Most intelligent irrigation products incur high installation costs that farmers cannot afford. CAIS used a low-cost component to achieve the same functionality that is found in intelligent irrigation products. Most low-cost irrigation systems water the plants from a single point that may lead to variable soil moisture if the terrain or the soil density is uneven. CAIS divided the experimental plot into multiple virtual planting areas (VPAs) and dynamically adapted the watering interval of each VPA to balance the soil moisture of the whole experimental plot. Results showed that the forecasting error of the SMF model was less than 12 moisture levels over a scale of 1024 levels. Furthermore, CAIS maintained the soil moisture of the whole experimental plot at a constant level within ±5 error points with multiple watering points.