ARTIFICIAL INTELLIGENCE EVENT-TRIGGERED CONTROL SYSTEM FOR REAL-TIME MONITORING AND PREDICTION OF GREENHOUSE CONDITIONS

Greenhouse farming is often limited by unstable environmental conditions and inefficient manual control methods. The paper designs an Artificial Intelligence Event-Triggered Control System (AIETCS) of real-time monitoring and prediction of greenhouse conditions using an Artificial Neural Network (ANN). A simulation based, quantitative methodology was embraced. Important environmental parameters were defined in terms of temperature, relative humidity, soil moisture, photo synthetically active radiation (PAR) and soil nutrients, between rainy and Harmattan seasons (2021-2024). Hardware interactions were modelled using Proteus simulation and data were pre-processed then used to train the ANN. The ANN was used to predict short-term greenhouse conditions and the event-triggered logic only used actuators when they were above their thresholds. Results show the ANN achieved high predictive accuracy (validation loss = 2.0761e-04, MAE = 0.028, R2=0.97R^2 = 0.97R2=0.97). Compared to baseline conditions, where deviations from ideal values reached 58% (soil moisture) and 39% (humidity), the AIETCS consistently reduced deviations to below 3.5% across all variables. The findings demonstrate that ANN-driven event-triggered control improves greenhouse stability, reduces resource wastage, and provides a scalable framework for precision agriculture in resource-constrained settings.