Scopus Indexed Publications

Paper Details


Title
A Data-Driven Approach to Sustainable Agriculture: Integrated IoT, Machine Learning, and Automation in Smart Greenhouses

Author
, Md. Moniruzzaman Hemal,

Email

Abstract

Meeting rising food demand sustainably requires production systems that increase yield per unit of water and energy while limiting environmental impact. We develop and evaluate an integrated controlled-environment agriculture platform that unifies horizontal greenhouse beds, vertical racks, and hydroponic channels with real-time sensing of temperature, humidity, light, and moisture; closed-loop climate and irrigation control; and a machine-learning crop recommender. An ESP32 telemetry layer streams data to the cloud for monitoring and analysis, while a dual-axis solar-tracking photovoltaic subsystem and automated rainwater harvesting reduce dependence on the grid and freshwater extraction. Experiments with a functional prototype deployed in Gazipur, Bangladesh, demonstrate stable closed-loop operation. The solar tracking system improved energy harvesting efficiency by approximately 18 % compared to static panels, while the rainwater system successfully captured precipitation to offset irrigation needs. The boosting ensemble achieved 97 % across accuracy, precision, recall, F1, and MCC, macro averaged using a one-vs-rest protocol. Detailed analysis reveals that while the model generalizes well, specific legume classes exhibit minor confusion due to similar nutrient profiles. These results show that tightly integrating perception, learning, and actuation within a resource-aware greenhouse can deliver high reliability and sustainability. Future work will expand local data collection to bridge the domain gap between public datasets and local sensor noise.


Keywords

Journal or Conference Name
IEEE International WIE Conference on Electrical and Computer Engineering/WIECON-ECE

Publication Year
2025

Indexing
scopus