Improving the Performance Against Force Variation of EMG Controlled Multifunctional Upper-Limb Prostheses for
Summary
This study introduces new electromyogram (EMG) features for robust control of prosthetic hands in amputees. These features improve classification accuracy by reducing the impact of varying force levels during prosthesis use.
Area of Science:
- Biomedical Engineering
- Rehabilitation Technology
- Signal Processing
Background:
- Robust control of prosthetic hands is crucial for transradial amputees.
- Variations in force levels significantly impact electromyogram (EMG) signal-based prosthesis control.
- Existing feature extraction methods struggle with force variability.
Purpose of the Study:
- To develop novel EMG features that enhance prosthesis control robustness against force variations.
- To reduce the impact of force level variations on EMG-controlled prostheses.
- To improve classification accuracy in prosthetic hand control systems.
Main Methods:
- Proposed a novel feature set characterizing EMG activity using spectral moments orientation.
- Employed a time-domain processing approach for efficient feature extraction.
- Evaluated features on EMG data from nine transradial amputees across six movement classes and three force levels.
Main Results:
- The proposed features demonstrated significant reductions in classification error rates.
- Achieved an average improvement of 6% to 8% in classification performance compared to existing methods.
- The method showed effectiveness across all subjects and force levels when trained with combined force data.
Conclusions:
- The novel EMG features effectively mitigate the impact of force variations in prosthetic hand control.
- This approach offers a more robust and computationally efficient solution for EMG-based prostheses.
- The findings contribute to improved functional outcomes for transradial amputees using prosthetic devices.
