A Minimal and Multi-Source Recording Setup for Ankle Joint Kinematics Estimation During Walking Using Only Proximal...

Area of Science:

• Biomechanics
• Biomedical Engineering
• Rehabilitation Technology

Background:

• Accurate ankle joint kinematics are crucial for prosthetic control and rehabilitation.
• Existing methods often require extensive sensor placement, impacting user comfort and practicality.
• Developing minimal sensor setups is key for improving prosthetic functionality and user experience.

Purpose of the Study:

• To propose and validate a minimal setup for estimating ankle joint kinematics.
• To investigate the use of proximal lower-limb information (thigh muscle activity and hip kinematics) for ankle motion prediction.
• To assess the feasibility of this approach for transfemoral amputee applications.

Main Methods:

• Surface electromyography (sEMG) recorded thigh muscle activity (Rectus Femoris, Biceps Femoris, Vastus Medialis).
• Hip and ankle joint kinematics were synchronously measured during unconstrained walking.
• Time-domain (TD) and wavelet (WT) features from sEMG data fed Artificial Neural Networks, Random Forest, and Least Squares-Support Vector Machine (LS-SVM) models.
• Model performance evaluated using Root Mean Square Error (RMSE).

Main Results:

• The Biceps Femoris muscle combined with LS-SVM achieved the best ankle angle estimation (RMSE < 5.6 degrees) using both TD and WT features.
• Incorporating hip joint trajectory significantly improved regression performance (RMSE < 4.5 degrees).
• Demonstrated the feasibility of estimating ankle kinematics with limited proximal lower-limb data.

Conclusions:

• A minimal setup using proximal lower-limb information can reliably estimate ankle joint kinematics.
• This approach enhances comfort and practicality for transfemoral amputee users.
• Findings support the development of data-driven models for active prosthetic control and rehabilitation solutions.