A Relationship Model between Optimized Exoskeleton Assistance and Gait Conditions Improves Multi-gait...

Area of Science:

• Biomechanical Engineering
• Human-Robot Interaction
• Rehabilitation Robotics

Background:

• Exoskeletons can reduce human walking metabolic cost, but personalized assistance is challenging due to individual variability and diverse gait conditions.
• Human-in-the-loop (HIL) optimization offers tailored assistance but often involves lengthy optimization periods.
• Efficient and effective exoskeleton assistance customization is crucial for widespread adoption and therapeutic benefits.

Purpose of the Study:

• To establish a predictive model for exoskeleton assistance parameters based on gait conditions (speed, slope).
• To validate the model's ability to accurately predict optimal assistance torque.
• To assess the impact of model-assisted optimization on metabolic cost and muscle activity.

Main Methods:

• Conducted a series of Human-in-the-loop (HIL) optimization experiments across various walking speeds and slopes.
• Developed a model to capture the relationship between gait conditions and optimized exoskeleton assistance parameters.
• Validated the model by comparing its predicted assistance torque with HIL-optimized values and evaluating metabolic and muscle activity reductions.

Main Results:

• The model-predicted assistance torque closely matched HIL-optimized torque.
• Model-calculated assistance reduced metabolic cost by 11.95% (p<0.001) and soleus activity by 22.28% (p=0.049).
• Model-initialized optimization yielded greater metabolic and muscle activity reductions compared to empirical values, improving efficiency by 50%.

Conclusions:

• A gait-condition-based model significantly enhances the efficiency and effectiveness of exoskeleton assistance customization.
• The model facilitates faster and more accurate tailoring of assistance for diverse walking scenarios.
• This approach expands the applicability of HIL optimization and improves exoskeleton performance for users.