How to Improve Fault Tolerance in Disaster Predictions: A Case Study about Flash Floods Using IoT, ML and Real Data

Area of Science:

• Environmental Science
• Computer Science
• Disaster Management

Background:

• Natural disasters pose a significant global threat, with urban areas facing amplified consequences.
• Traditional methods for disaster management rely on wireless sensor networks (WSNs) and machine learning (ML).
• Emerging technologies, including IP-based sensor networks and the Internet of Things (IoT), offer new avenues for environmental monitoring and forecasting.

Purpose of the Study:

• To introduce and evaluate SENDI (System for dEtecting and forecasting Natural Disasters based on IoT).
• To demonstrate a fault-tolerant system for natural disaster detection, forecasting, and alert issuance.
• To showcase the system's capability in extreme situations and its application in flash flood forecasting.

Main Methods:

• Development of the SENDI system, integrating IoT, ML, and WSN technologies.
• System modeling using ns-3 network simulator.
• Utilizing real-world data from a WSN deployed in São Carlos, Brazil, for river monitoring.
• Implementing fault-tolerance mechanisms to address communication breakdowns and node destruction.

Main Results:

• SENDI demonstrates effective detection and forecasting of natural disasters.
• The system maintains functionality and data distribution even under extreme conditions.
• A case study on flash flood forecasting using the SENDI model and WSN data shows promising results.

Conclusions:

• SENDI offers a robust and fault-tolerant solution for natural disaster management.
• The integration of IoT, ML, and WSN enhances environmental monitoring and disaster preparedness.
• The system shows potential for real-world application in urban disaster mitigation and early warning systems.