Human Gait Modeling, Prediction and Classification for Level Walking Using Harmonic Models Derived from a Single...

Area of Science:

• Biomechanics
• Robotics
• Human Motion Analysis

Background:

• Current human gait modeling relies on hip, foot, or thigh acceleration, often yielding suboptimal regeneration accuracy.
• Accurate gait modeling is crucial for applications like prosthetics, robotics, and rehabilitation.

Purpose of the Study:

• To present a new harmonic approach for modeling human gait during level walking.
• To demonstrate the effectiveness of using gyroscopic signals and thigh flexion-extension from a single thigh-mounted Inertial Measurement Unit (IMU).

Main Methods:

• Utilized a single thigh-mounted IMU to capture gyroscopic signals and thigh flexion-extension data.
• Applied harmonic analysis to model thigh angle and gyro signals, determining the number of significant harmonics required for accurate regeneration.
• Estimated the fundamental frequency of the harmonic model using stride time.

Main Results:

• Thigh angle was accurately modeled using five significant harmonics (correlation > 0.999, RMSE < 0.5°).
• Gyro signals achieved comparable regeneration accuracy with nine significant harmonics.
• The fundamental frequency was estimated with a low error rate of 0.0479% (±0.0029%).
• Six common stride patterns and their corresponding harmonic models were presented.

Conclusions:

• Human gait in level walking can be effectively modeled using harmonic analysis of thigh angle or gyro signals from a single thigh-mounted IMU.
• This harmonic approach offers higher regeneration accuracy than existing gait modeling techniques.
• The developed harmonic models show potential for stride prediction, classification, and activity recognition.