Researchers affiliated with Mansoura National University and Kafrelsheikh University published a study in Scientific Reports on 12 October modelling a smart-irrigation system for Egypt’s New Delta that combines AI, an IoT architecture and solar power in MATLAB. The team used NASA POWER data for environmental inputs. The design links LSTM-based irrigation-pump state classification with photovoltaic generation, battery state-of-charge management, maximum-power-point tracking and adaptive pump-speed control.

The paper reports 96.11% accuracy for classifying pump operating states with the LSTM model. Against an assumed fixed-schedule irrigation baseline, the simulation estimated 28.57% annual water savings; adaptive pump-speed control reduced modeled motor energy use by 74.74%. These are outputs of a MATLAB scenario using real environmental inputs, not savings measured with sensors in a field. The authors identify hardware implementation and long-term real-world trials as future work.