Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Abstract. Irrigation scheduling is of vital importance in the agricultural irrigation management for higher water use efficiency in the context of water scarce. Among various methods committed to obtaining more precise irrigation quantity with optimal timing at daily scale, model-based methods were announced to be more efficient due to their ability to integrate the atmosphere-plant-soil effects on irrigation scheduling. However, these models simulating plant growth under different irrigation treatment were hardly employed in real-time irrigation decision support system, either because of their expertise in modeling or the high cost of constructing such systems. This study used a neural network framework to predict the soil moisture across the root zone of crops, thereby to extend the application of water stress-based irrigation scheduling method based on Root Zone Water Quality Model. The trained NN model was then used to predict soil moisture and generate irrigation schedules. Furthermore, the trained NN model was embedded into a microcomputer, Raspberry-Pi (RPi), to process the control signal to irrigation valves of a drip irrigation system. Only weather data was required as the input of RPi. The RPi estimates root zone soil moisture daily and triggers irrigation automatically. An irrigation decision support system was developed consisting of RPi as the decision maker, data collect system, communication module and valves. This study provides an example of applying artificial intelligent algorithms in real-time irrigation scheduling decision support system for higher water use efficiency and lower cost.