KURTAR — IoT & AI-Based Disaster Victim Detection and Coordination Platform
Jan 1, 2024
·
2 min read
Project Overview
The KURTAR Project aims to revolutionize disaster response by combining IoT, mobile devices, and AI-driven decision systems. During earthquakes or similar emergencies, the platform collects real-time data from mobile phones, wearables, and distributed IoT sensors. This data is processed by machine learning and deep learning models, enabling autonomous decision-making to assist rescue teams in coordination and victim detection.
Key Features
- Multi-Sensor Fusion: Integration of accelerometer, gyroscope, GPS, and audio signals from mobile and IoT devices.
- Edge/Cloud AI: Lightweight models (CNN, LSTM, CRNN, GNN) deployed on Raspberry Pi, mobile, and cloud servers for real-time analysis.
- Crowdsensing: Victim smartphones and wearables act as distributed data points in the network.
- Dynamic Sensor Networks: Self-organizing wireless sensor networks for resilient communication in disaster zones.
- Decision Support: Bayesian networks and reinforcement learning models support routing, prioritization, and risk mapping for rescue teams.
- Mapping & Visualization: GIS-based dashboards provide live situational awareness for AFAD and emergency responders.
Research Context
The project aligns with my doctoral research:
- Artificial Intelligence-Supported Real-Time Disaster Management with Multi-Sensor Fusion (PhD, Kocaeli University, ongoing).
- Supported by a TÜBİTAK 1001 Grant, integrating academic research with practical disaster-response applications.
Impact
KURTAR has the potential to save lives in the golden hours of disaster response, by accelerating victim detection, reducing coordination delays, and enabling data-driven decision-making for national and international rescue teams.
