Gait analysis using gravitational acceleration measured by wearable sensors

Area of Science:

• Biomechanics
• Wearable Technology
• Human Motion Analysis

Background:

• Accurate human gait analysis is crucial for diagnosing various musculoskeletal and neurological conditions.
• Existing methods for gait measurement often rely on complex laboratory setups or have limitations in capturing dynamic, real-world movements.
• Wearable sensors offer a promising approach for unobtrusive and continuous gait monitoring.

Purpose of the Study:

• To propose and validate a novel method for measuring human gait posture using wearable sensor units.
• To develop an algorithm capable of accurately estimating joint angles and 3D positions during walking.
• To provide quantitative data for improved gait diagnosis.

Main Methods:

• Utilized seven wearable sensor units, each equipped with tri-axial accelerometers and gyroscopes, placed on the abdomen and lower limb segments.
• Developed an optimization analysis to isolate gravitational acceleration from translational acceleration and noise in sensor data.
• Employed Fast Fourier Transform (FFT) analysis to identify characteristic frequencies for gravitational acceleration pattern assumption.
• Calculated joint positions using segment lengths, estimated joint angles, and physiological motion constraints, solving an inverse problem for optimized gravitational acceleration patterns.

Main Results:

• Successfully measured simultaneous acceleration data from all sensor segments during walking.
• Visualized characteristic 3D gait patterns using a stick figure model.
• Enabled visualization of knee joint trajectories in the horizontal plane on a PC.
• Demonstrated the feasibility of the method with measurements from three healthy volunteers.

Conclusions:

• The proposed wearable sensor system and analysis method provide accurate quantitative information for human gait posture measurement.
• The technique effectively separates gravitational acceleration, enabling precise 3D joint position estimation.
• This novel approach holds significant potential for clinical gait diagnosis and rehabilitation monitoring.