This paper presents the design and deployment of an IoT-based early warning system for forest fire detection, optimized for remote and infrastructure-scarce environments. The system integrates multiple environmental sensorsincluding air temperature, air humidity, soil moisture, rainfall, wind speed, and wind directioninterfaced with ESP32 microcontrollers for local data acquisition. Each sensor node operates autonomously using solar power, enhancing sustainability and long-term field operation. Data is transmitted using LoRaWAN for long-range, low-power communication, with 3G fallback for redundancy in critical zones. A centralized server collects and processes the data in real time, applying a rule-based threshold algorithm combined with temporal trend analysis to detect fire-prone conditions. A web-based dashboard provides monitoring and visualization, while alert notifications are dispatched via SMS and email when risk levels exceed defined thresholds. Field trials conducted in the An Giang region demonstrated reliable communication over distances exceeding 5 km, stable solar-powered operation, and alert latency of under 10 seconds. The results confirm the system’s feasibility and effectiveness in providing early warnings for wildfire prevention in rural forest areas.
Introduction
Vietnam prioritizes forest conservation but faces frequent wildfires caused by dry weather and climate change. Accurate and timely forest fire risk assessment requires continuous monitoring of environmental data such as temperature, humidity, rainfall, and wind. Traditional data collection methods are labor-intensive and often outdated, prompting the development of an IoT-based automated warning system.
The proposed system uses real-time sensors and wireless communication technologies to monitor forests continuously. Two communication models were tested: one based on 3G mobile networks powered by solar energy, and another using LoRaWAN (Long Range Wide Area Network) with solar power. The LoRaWAN model proved more suitable for remote, infrastructure-poor forest areas due to its long-range, low-power capabilities.
Data from sensors measuring temperature, humidity, rainfall, wind speed, and soil moisture are transmitted to a central server via either network. The server processes the data to calculate a meteorological fire index, classifying fire risk into five levels. The system provides real-time fire risk alerts accessible through a web platform, enabling forest managers and the public to respond promptly.
The study also reviews related research on IoT applications for wildfire detection globally, highlighting the importance of combining sensor networks, climate models, and sustainable power solutions like solar energy. The system’s architecture ensures energy efficiency, scalability, and reliable communication, especially with LoRaWAN technology, which is ideal for monitoring large, remote forest areas.
Conclusion
This study proposed and experimentally deployed two IoT-based solutions for forest fire early warning in Tinh Bien district, An Giang province: one using LoRaWAN technology and another using 3G-based communication. Both systems operated reliably and delivered real-time data and fire risk estimates. However, the 3G-based model proved more appropriate given the current local infrastructure, particularly in terms of internet access and electricity availability. The system performed autonomous data acquisition every 5–7 minutes, continuously throughout the day, providing accurate daily calculations of the fire index (P value). Data was uploaded to a cloud-based web dashboard, enabling forest rangers and management agencies to monitor risk levels conveniently, anytime and anywhere.
References
[1] X. Zong, X. Tian, and Y. Yin, \"Impacts of Climate Change on Wildfires in Central Asia,\" Forests, vol. 11, no. 8, pp. 802, 2020. doi: 10.3390/f11080802.
[2] E. Sáez, J. R. Sánchez, M. G. Parra, and P. Hernández, \"IoT-based wildfire detection system,\" Sensors, vol. 19, no. 4, pp. 876, 2019. doi: 10.3390/s19040876.
[3] S. Ghosh, M. Chatterjee, and A. P. Singh, \"Wireless sensor networks for forest fire monitoring,\" Journal of Computer Networks and Communications, vol. 2017, pp. 1-12, 2017. doi: 10.1155/2017/7120378.
[4] A. Carvalho, S. T. Silva, and F. T. Lima, \"Fire risk prediction models using environmental and climatic data,\" Environmental Modelling& Software, vol. 98, pp. 16-27, 2017. doi: 10.1016/j.envsoft.2017.09.017.
[5] N. Rani, S. R. Reddy, and G. S. R. K. Rao, \"Solar-powered IoT-based forest fire monitoring system,\" IEEE Access, vol. 7, pp. 40254-40263, 2019. doi: 10.1109/ACCESS.2019.2909567.
[6] A. Firoz, M. A. Hossain, and M. R. Sarker, \"Challenges in forest fire detection using IoT-based systems,\" International Journal of Computer Applications, vol. 181, no. 28, pp. 19-26, 2019. doi: 10.5120/ijca2019918653.