Online electromyographic control of a robotic prosthesis
Pradeep Shenoy1, Kai J Miller, Beau Crawford
1Department of Computer Science and Engineering, University of Washington, P.O. Box 352350, Seattle, WA 98195, USA. pshenoy@cs.washington.edu
IEEE Transactions on Bio-Medical Engineering
|March 13, 2008
Summary
This study demonstrates accurate robotic arm control using forearm surface electromyography (EMG) signals. Real-time EMG control achieved high accuracy with minimal training, outperforming keyboard control in complex tasks.
Area of Science:
- Biomedical Engineering
- Robotics
- Neuroscience
Background:
- Surface electromyography (EMG) is a non-invasive method to measure electrical activity produced by skeletal muscles.
- Myoelectric control, utilizing EMG signals, offers a promising avenue for prosthetic and robotic limb control.
- Existing classification-based paradigms for myoelectric control require further investigation for enhanced real-time applications.
Purpose of the Study:
- To investigate the efficacy of forearm surface EMG signals for real-time robotic arm control.
- To extend current classification-based myoelectric control paradigms for improved accuracy and speed.
- To design and evaluate an online EMG-based control system for a 4 degrees of freedom robotic arm.
Main Methods:
- An offline study was conducted to classify eight distinct EMG signal classes, achieving high accuracy (92-98%) at 16 classifications/s.
- A real-time online control system for a 4 degrees of freedom robotic arm was designed using EMG signals.
- A three-subject study compared the performance of the EMG-based system against a keyboard-control baseline for various complex tasks.
Main Results:
- The offline classification achieved high accuracy (92-98%) with minimal training time (under 10 minutes).
- The real-time EMG-based robotic arm control system demonstrated effective performance in complex tasks.
- Comparison with keyboard control indicated the potential of EMG for intuitive and efficient robotic arm manipulation.
Conclusions:
- Forearm surface EMG signals can achieve high-accuracy myoelectric control of robotic arms.
- The developed EMG control system requires minimal user training, making it practical for real-world applications.
- EMG-based control presents a viable and potentially superior alternative to traditional control methods like keyboard interfaces for robotic systems.
