Human-in-the-Loop Control Framework for Robot-Mediated Error Augmentation Training Based on Muscle Synergy...

Area of Science:

• Robotics
• Neuroscience
• Biomechanics

Background:

• Robot-mediated error augmentation (EA) shows promise for motor adaptation.
• Integrating EA into human-in-the-loop (HITL) control is difficult due to a lack of adaptation metrics and personalized control.
• Existing methods struggle to adapt EA intensity to individual motor strategies.

Purpose of the Study:

• To propose a novel HITL control framework for personalizing EA intensity.
• To leverage muscle synergy analysis for guiding EA personalization.
• To develop a synergy-based performance metric for quantifying motor adaptation.

Main Methods:

• Developed a HITL control framework integrating muscle synergy analysis.
• Designed a synergy-based performance metric measuring structural similarity across muscle synergies.
• Employed a Gaussian process-based Bayesian algorithm to adapt EA gain based on individual responses.
• Validated the framework in an 8-day study with ten healthy subjects performing reaching tasks.

Main Results:

• The experimental group showed significantly greater reductions in trajectory deviation compared to the control group.
• Robust within-group improvements (p < 0.01) and positive inter-group trends in muscle synergy similarities were observed.
• Demonstrated reliable convergence of the optimization process and robust inter-subject adaptability.
• The synergy metric effectively captured motor adaptation.

Conclusions:

• The proposed HITL framework effectively personalizes EA intensity using muscle synergy analysis.
• This approach significantly enhances motor adaptation and trajectory control.
• The findings support the framework's potential as a foundational tool for adaptive motor training in human-robot interaction.