Smart Seismic Sensing for Indoor Fall Detection, Location, and Notification.

Area of Science:

• Engineering
• Computer Science
• Gerontology

Background:

• Falls are a significant risk for elderly individuals living alone, necessitating advanced monitoring solutions.
• Existing fall detection systems often rely on wearable sensors or complex vision-based methods.
• There is a need for non-intrusive, accurate, and real-time fall detection and identification systems.

Purpose of the Study:

• To develop and evaluate a novel real-time smart system for fall detection, location, and notification using only floor vibration data.
• To incorporate person identification based on footstep vibrations for enhanced emergency response.
• To improve the robustness and accuracy of fall detection against similar events and identify individuals with minimal data.

Main Methods:

• Utilizing floor vibration data as the sole source for fall recognition and person identification.
• Implementing a collaborative in-network localization method with sensor nodes.
• Employing a voting system among sensor nodes to enhance person identification accuracy.
• Developing algorithms to distinguish fall events from non-fall events like jumps or object falls.

Main Results:

• The system achieved a fall detection acceptance rate of 95.14%, effectively distinguishing falls from other events.
• Person identification accuracy reached 97.22% using only one or two footsteps, outperforming existing methods.
• The fall down location error was less than 0.27 meters, providing precise localization.

Conclusions:

• The developed floor vibration-based system offers a robust and accurate solution for real-time fall detection and person identification.
• This technology significantly advances smart home safety, particularly for elderly individuals living independently.
• The system's ability to integrate with commercial smart devices enables immediate emergency notifications, improving response times and safety.