A framework for detecting unfolding emergencies using humans as sensors.

Area of Science:

• Computer Science
• Social Computing
• Crisis Informatics

Background:

• Online social networks (OSNs) and smartphones enable social sensing systems.
• Human sensing capabilities are increasingly used for emergency detection and crisis management.
• Existing systems face challenges in capturing, filtering, and extracting meaningful data from OSNs.

Purpose of the Study:

• To propose a conceptual and architectural framework for emergency detection systems.
• To define an ontology for the
• human as a sensor
• (HaaS) paradigm in emergency detection.
• To design a modular, type-independent architecture for emergency detection.

Main Methods:

• Developed a conceptual and architectural framework for HaaS-based emergency detection.
• Defined an ontology for the HaaS paradigm.
• Designed a modular architecture adaptable to various emergency types.
• Implemented an earthquake detection application using Twitter data.

Main Results:

• The proposed architecture demonstrated effectiveness in detecting earthquakes.
• Experimental results validated the system's performance using real-world Twitter data from Italy.
• The framework successfully captured, filtered, and extracted relevant information for crisis management.

Conclusions:

• The HaaS paradigm offers a viable approach for developing robust emergency detection systems.
• The proposed modular architecture is versatile and can be applied to different emergency scenarios.
• Social sensing systems, leveraging human input, significantly enhance crisis management capabilities.