Electromyography pattern-recognition based prosthetic limb control using various machine learning techniques.

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Machine Learning Applications

Background:

• Upper extremity amputation affects a significant portion of the population, impacting daily life and independence.
• Current prosthetic limb control relies on muscle contractions, lacking precise force feedback for object manipulation.
• Existing methods may require invasive surgical procedures for prosthetic integration.

Purpose of the Study:

• To develop a non-invasive method for predicting the required force for prosthetic limb control.
• To enhance the self-reliance and quality of life for individuals with amputations.
• To utilize machine learning regression models for accurate force prediction in prosthetic applications.

Main Methods:

• Implementation of Machine Learning (ML) regression models, including Support Vector Regressor (SVR), Linear Regression, and Random Forest.
• Training and evaluation of models to predict the force needed to regulate voltage for servomotors in prosthetic limbs.
• Comparative analysis of model performance for force requirement prediction.

Main Results:

• The Random Forest model achieved highly accurate predictions for the force required to control servomotor voltage.
• All tested ML models showed potential in predicting force requirements for prosthetic control.
• The study highlights the efficacy of ML in improving prosthetic functionality.

Conclusions:

• Machine learning, particularly the Random Forest model, offers a promising solution for precise force control in prosthetic limbs.
• This approach can mitigate the limitations of current prosthetic control systems by providing better force regulation.
• The developed method has the potential to significantly improve the user experience and functional capabilities of prosthetic devices.