Hebishy et al. (2026) Low-cost IoT real-time environmental monitoring and one-hour LSTM forecasting for small-scale agricultural applications

Identification

• Journal: Scientific Reports
• Year: 2026
• Date: 2026-09-15
• Authors: Ganna H. Hebishy, Ahmed K. Hassan, Samaa F. Osman, Mohamed S. Saraya
• DOI: 10.1038/s41598-026-69489-0

Research Groups

Mechatronics Engineering Department, Faculty of Engineering, Mansoura University, Egypt. Assistant Professor of Computers Engineering and Control Systems, Faculty of Engineering, Mansoura National University, Egypt. Electrical Engineering Department, Faculty of Engineering, Mansoura University, Egypt. Computers Engineering and Control Systems Department, Faculty of Engineering, Mansoura University, Egypt.

Short Summary

This study introduces a low-cost IoT system for environmental monitoring and forecasting in small-scale agricultural applications. The system was deployed at a residential garden in the Nile Delta region of Egypt and utilized univariate LSTM models to forecast one hour ahead for each of seven sensor channels.

Objective

• Investigate the technical feasibility of pairing affordable hardware with deep learning forecasting for continuous environmental monitoring.
• Develop an IoT system capable of tracking temperature, humidity, precipitation, soil moisture, methane, carbon monoxide, and air quality in real-time.

Study Configuration

• Spatial Scale: Small-scale agricultural application at a residential garden site in the Nile Delta region of Egypt.
• Temporal Scale: 21-day deployment period with data collected every minute.

Methodology and Data

• Models used: Univariate Long Short-Term Memory (LSTM) networks.
• Data sources: Sensor data from seven channels: temperature, humidity, precipitation, soil moisture, methane, carbon monoxide, and air quality.

Main Results

• The system captured 31,460 raw observations across the 21-day window, with 30,981 retained after preprocessing (98.48% of the exported record).
• Distinct environmental patterns were observed, including a rising trend of 0.296 °C/day in temperature and an anti-phase coupling between temperature and humidity.
• The LSTM models showed modest performance improvements over persistence for methane and air quality forecasting.

Contributions

• This study demonstrates the technical feasibility of pairing affordable hardware with deep learning forecasting for continuous environmental monitoring.
• The system provides a foundation for future irrigation-support applications in resource-constrained settings.

Funding

• This research was funded by Mansoura University, Egypt.

Citation

@article{Hebishy2026Lowcost,
  author = {Hebishy, Ganna H. and Hassan, Ahmed K. and Osman, Samaa F. and Saraya, Mohamed S.},
  title = {Low-cost IoT real-time environmental monitoring and one-hour LSTM forecasting for small-scale agricultural applications},
  journal = {Scientific Reports},
  year = {2026},
  doi = {10.1038/s41598-026-69489-0},
  url = {https://doi.org/10.1038/s41598-026-69489-0}
}