Adaptive model-based assistive control for pneumatic direct driven soft rehabilitation robots.

Area of Science:

• Robotics
• Rehabilitation Engineering
• Biomedical Engineering

Background:

• Effective robot-assisted therapy requires inherent compliance and assistive behavior.
• Soft rehabilitation robots offer inherent compliance for safe human-robot interaction.
• Personalized control is crucial for relearning motor functions and muscle reconstruction.

Purpose of the Study:

• To present two adaptive model-based assistive controllers for pneumatic soft rehabilitation robots.
• To enable controllers to adapt to individual patient behavior, effort, and ability.
• To enhance robot transparency and provide tailored assistance during rehabilitation.

Main Methods:

• Development of two adaptive model-based assistive controllers.
• Utilizing separated models of the soft robot and the patient's extremity.
• Employing an inverse model of the soft robot and inverse adaptive models of the patient's extremity.
• Online learning of individual patient behavior and effort.

Main Results:

• The controllers demonstrated effectiveness in assisting elbow training with unimpaired subjects.
• The adaptive models allowed for personalized assistance based on patient needs.
• Robot transparency was achieved through an inverse model of the soft robot.

Conclusions:

• Adaptive model-based control is effective for soft rehabilitation robots.
• Personalized assistance is essential for optimizing rehabilitation outcomes.
• The developed controllers show promise for neurological and orthopedic rehabilitation.