AGRICULTURAL WATER MANAGEMENT USING ARTIFICIAL NEURAL NETWORK

Abstract

This work presents the development of an artificial intelligence (AI) system for efficient management of agricultural water resources using Artificial Neural Network (ANN). The current available technology required specific calibration and is costly. In this work, a low cost soil moisture sensor was used to determine soil moisture percentage while a prediction model was generated by using the ANN to experimentally estimate both the in-situ water volume and probing depth. The result showed regression value (R) and Mean Square Error (MSE) of 0.9929, and 0.4972, and 0.9947, and 0.1789, respectively, for loam and peat soil samples. Experimental investigation on test samples revealed a mean accuracy of 80%, and this study showed promising performance of the developed ANN model for agricultural applications. In the future, an integration of the developed system with a feedback water supply system would be beneficial for use in agricultural water management.