Scopus Indexed Publications

Paper Details


Title
Design and Implementation of a LoRa-Based Flood Monitoring System for Real-Time Disaster Management and Early Warning

Author
, Fayazunnesa Chowdhury,

Email

Abstract

Bangladesh faces daunting issues, including flooding from 907 rivers and 25% of the country at risk of annual coastal flooding by 2050. Conventional flood monitoring systems are challenged by a lack of real-time monitoring efficiency, high unit costs, and difficulty in rural networking. This work presents a LoRa-based IoT Flood Monitoring System for real-time disaster management in flood-prone areas. The system integrates multiple sensors (ultrasonic water level, rainfall, flow rate, temperature, and humidity) powered by solar energy and ESP32 microcontrollers, achieving 10–12 km of data transmission without cellular service. Two complementary evaluations were performed: (1) a synthetic dataset study based on historical flood records to assess detection accuracy, and (2) a 3-month rural Bangladesh field deployment to validate network performance. Synthetic dataset experiments (with ablation studies) achieved 94.2% accuracy, 91.1% precision, and 97.2% recall in flood detection, while ensemble learning achieved 94.1% accuracy in 10-fold cross-validation. The field trial demonstrated 97% packet delivery, 95 ms latency, and 18 s alert delivery to rangers. Performance remained robust in adverse weather, maintaining over 89% packet success rate during heavy rainfall and electromagnetic interference. This research presents a sustainable, low-cost flood monitoring solution suitable for developing regions with limited infrastructure, with demonstrated potential for global scalability.


Keywords

Journal or Conference Name
Procedia Computer Science

Publication Year
2026

Indexing
scopus