Assist-as-needed Hip Exoskeleton Control for Gait Asymmetry Correction via Human-in-the-loop Optimization.

Area of Science:

• Biomedical Engineering
• Rehabilitation Robotics
• Neuroscience

Background:

• Gait asymmetry is a common issue for stroke survivors, impacting mobility and quality of life.
• Existing robot-aided gait rehabilitation lacks focus on correcting hemiplegic gait asymmetry.
• Controlling exoskeletons for patients with partial motor control presents significant challenges in shared control.

Purpose of the Study:

• To propose an assist-as-needed (AAN) hip exoskeleton control system.
• To optimize control for correcting gait asymmetry in hemiplegic gait using human-in-the-loop.
• To enhance active patient participation and maximize rehabilitation efficiency.

Main Methods:

• Developed an objective function for real-time gait performance and participation evaluation.
• Implemented online tuning of control parameters based on subject-specific data.
• Utilized human-in-the-loop optimization for exoskeleton control.
• Conducted experimental studies with simulated hemiplegic gait.

Main Results:

• The proposed AAN control effectively stimulated greater subject contribution to movement.
• Online parameter tuning adapted to individual movement variability.
• The system demonstrated feasibility in correcting gait asymmetry.
• Validated the hypothesis that AAN controls enhance active human participation from a biomechanical perspective.

Conclusions:

• The AAN hip exoskeleton control is a feasible approach for hemiplegic gait rehabilitation.
• Human-in-the-loop optimization significantly improves active participation in stroke survivors.
• This method offers a promising direction for personalized and effective gait rehabilitation.