Fall Detection System Based on Point Cloud Enhancement Model for 24 GHz FMCW Radar

Area of Science:

• Engineering
• Computer Science
• Gerontology

Background:

• Automatic fall detection is crucial for senior citizen health monitoring.
• Millimeter-wave radar offers privacy-preserving, cost-effective human pose recognition.
• Low-quality radar data challenges reliable fall detection.

Purpose of the Study:

• To develop a low-cost model for high-quality 3D human point cloud generation.
• To enhance the accuracy and effectiveness of automatic fall detection systems.
• To address the limitations of current fall detection methods using millimeter-wave radar.

Main Methods:

• Proposed a novel model for generating high-quality 3D human point clouds from low-cost hardware.
• Developed a system extracting distribution features using small millimeter-wave radar antenna arrays.
• Utilized advanced signal processing techniques for point cloud enhancement.

Main Results:

• Achieved 99.1% accuracy for fall detection on new subjects.
• Attained 98.9% accuracy for fall detection in new environments.
• Demonstrated the system's effectiveness in improving point cloud quality and detection reliability.

Conclusions:

• The proposed low-cost millimeter-wave radar system significantly improves automatic fall detection accuracy.
• This technology offers a practical solution for enhanced senior health monitoring.
• The system's performance in diverse conditions highlights its robustness and potential for widespread adoption.