Dexterous Force Estimation during Finger Flexion and Extension Using Motor Unit Discharge Information

Area of Science:

• Biomedical Engineering
• Neuroscience
• Robotics

Background:

• Advanced robotic hands require reliable neural-machine interfaces for intuitive control.
• Accurate estimation of individual finger forces is crucial for dexterous robotic manipulation.

Purpose of the Study:

• To develop and validate a novel method for continuous, concurrent estimation of individual finger isometric forces.
• To improve upon conventional electromyogram (EMG) amplitude-based methods for force prediction.

Main Methods:

• Extracted motor unit (MU) discharge activity from high-density surface EMG signals of finger flexors and extensors.
• Grouped MU information to associate with individual finger flexion/extension.
• Utilized a linear regression model to predict individual finger forces during multi-finger tasks.

Main Results:

• The proposed MU-based method demonstrated superior force estimation performance compared to conventional EMG amplitude methods.
• Achieved higher correlation and lower estimation error between predicted and measured forces.
• Validated the method's effectiveness during dexterous multi-finger flexion and extension tasks.

Conclusions:

• The novel MU discharge activity-based method provides a more accurate approach to estimating individual finger forces.
• This technique shows potential for developing robust neural-machine interfaces for intuitive robotic hand control.
• Further research can enhance the application of this method in advanced prosthetics and robotics.