Gait Phase Detection Based on Muscle Deformation with Static Standing-Based Calibration.

Area of Science:

• Biomechanics
• Wearable Technology
• Machine Learning

Background:

• Gait phase detection is crucial for applications like robotic assistance and health monitoring.
• Existing systems often rely on inertial, electromyography, or force myography sensors.
• Calibration typically requires data collection during walking, which can be time-consuming.

Purpose of the Study:

• To develop a novel gait phase detection system utilizing muscle deformation.
• To implement a static standing-based calibration method that reduces setup time.
• To evaluate the accuracy and feasibility of the proposed system.

Main Methods:

• Utilized muscle deformation information for gait phase detection.
• Developed a static standing-based calibration procedure using various postures.
• Employed a logistic regression algorithm, adjusting output with sensor angular velocity.
• Conducted experiments with 10 subjects, validating results against video data.

Main Results:

• Achieved a median accuracy of approximately 90% for foot-contact and foot-off detection.
• Demonstrated successful gait phase detection after a short, 60-second calibration period.
• Validated the feasibility of static standing-based calibration for real-time gait analysis.

Conclusions:

• The proposed system offers a feasible and efficient method for gait phase detection.
• Static standing-based calibration using muscle deformation significantly simplifies the setup process.
• This technology has potential for improved wearable-based health monitoring and assistive device control.