Related Experiment Video
Updated: Apr 26, 2026
11:16
Engineering Platform and Experimental Protocol for Design and Evaluation of a Neurally-controlled Powered Transfemoral Prosthesis
Published on: July 22, 2014
15.6K
Source selection for real-time user intent recognition toward volitional control of artificial legs
IEEE Journal of Biomedical and Health Informatics
|July 24, 2014
Summary
Surface electromyography (EMG) signals and ground reaction forces are key for recognizing user intent in powered artificial legs. A select few data sources ensure accurate control, improving prosthetic leg functionality.
Area of Science:
- Biomedical Engineering
- Rehabilitation Robotics
- Human-Machine Interfaces
Background:
- Accurate user intent recognition is crucial for effective volitional control of powered artificial legs.
- A debate exists regarding the optimal data sources for precise and responsive prosthetic control.
Purpose of the Study:
- Investigate the utility of various data sources for user intent recognition.
- Identify an informative subset of data sources for prosthetic leg volitional control.
Main Methods:
- Collected surface electromyography (EMG) signals, ground reaction forces/moments, and kinematic data from transfemoral amputees.
- Employed three source selection algorithms to rank and select informative data sources.
- Evaluated system performance across two experimental days with four subjects.
Main Results:
- EMG signals and ground reaction forces/moments proved more informative than prosthesis kinematics.
- 9-11 data sources achieved 95% accuracy in recognizing seven tasks in real-time.
- Selected data sources demonstrated consistent performance, indicating robustness.
Conclusions:
- A reduced set of data sources, primarily EMG and ground reaction forces, is sufficient for accurate prosthetic leg control.
- Proposed a protocol for selecting informative data sources and sensor configurations for future powered prosthetic development.
