Ubiquitous Gait Analysis through Footstep-Induced Floor Vibrations

Area of Science:

• Biomechanics
• Sensor Technology
• Digital Health

Background:

• Quantitative gait analysis is crucial for diagnosing and monitoring neurological and musculoskeletal disorders.
• Traditional gait analysis methods (cameras, wearables) have practical limitations in daily life settings.
• There is a need for non-intrusive, privacy-friendly, and continuous gait monitoring solutions.

Purpose of the Study:

• To develop and validate a novel gait analysis system using passive floor vibration sensing.
• To estimate key temporal, spatial, and gait health parameters from floor vibrations.
• To address the challenge of floor type variability and develop adaptive algorithms.

Main Methods:

• Utilized vibration sensors mounted on the floor to passively capture footstep-generated vibrations.
• Developed floor-adaptive algorithms to generalize feature extraction across different surfaces.
• Conducted real-world walking experiments with 20 adults, analyzing over 12,000 gait cycles on concrete and wooden floors.

Main Results:

• Achieved high accuracy in estimating temporal parameters (90.5%, RMSE 0.08s).
• Demonstrated good accuracy for spatial parameters (71.3%, RMSE 0.38m).
• Showcased strong accuracy for gait health indicators (92.3%, RMSPE 7.7%).

Conclusions:

• Passive floor vibration sensing offers a viable, low-cost, and non-intrusive method for gait analysis.
• The developed floor-adaptive algorithms enable reliable gait parameter estimation in diverse environments.
• This technology holds promise for continuous, in-home gait health monitoring and early disease detection.