Analyzing Optimal Wearable Motion Sensor Placement for Accurate Classification of Fall Directions

Area of Science:

• Biomechanics
• Wearable Technology
• Gerontology

Background:

• Falls are a major health risk, especially for older adults.
• Accurate fall-direction classification is crucial for effective intervention.
• Previous research focused on fall detection, not direction classification across body regions.

Purpose of the Study:

• To determine optimal placement of wearable sensors for fall-direction classification.
• To compare the effectiveness of different sensor locations and body regions.
• To identify the best sensor configuration for accurate fall-direction analysis.

Main Methods:

• Assessed inertial measurement unit (IMU) sensors at 12 body locations.
• Compared various machine learning classifiers, including support vector machine (SVM).
• Evaluated sensor performance on left vs. right body sides and combined upper/lower body regions.

Main Results:

• Support vector machine (SVM) demonstrated superior performance across all sensor locations.
• No significant performance difference was found between left and right body-side sensor placements.
• Optimal sensor locations for fall-direction classification included the pelvis, upper legs, shoulder, and head.

Conclusions:

• The support vector machine (SVM) is the most effective classifier for fall-direction detection.
• Combining sensors from the pelvis, upper legs, and lower legs provides the optimal configuration for fall-direction classification.
• Strategic sensor placement is key for improving fall detection and prevention technologies.