Feedforward model based arm weight compensation with the rehabilitation robot ARMin

Area of Science:

• Robotics in Rehabilitation Medicine
• Neurorehabilitation Engineering
• Biomechanics of Human Movement

Background:

• Stroke survivors often experience severe upper limb motor deficits, limiting early-stage rehabilitation.
• Existing mechanical arm weight support systems can be restrictive and introduce unwanted disturbances.
• A need exists for adaptable, generalizable, and scalable arm weight compensation in robotic rehabilitation.

Purpose of the Study:

• To introduce a model-based feedforward weight compensation method for upper and lower arm rehabilitation.
• To develop a system that is online adjustable and generalizable across different robotic platforms.
• To validate the effectiveness of the proposed weight compensation in reducing patient effort during robotic-assisted training.

Main Methods:

• Development of a model-based feedforward control strategy for arm weight compensation.
• Integration and testing of the method with the ARMin V upper extremity rehabilitation robot.
• Validation of the system's performance using electromyography (EMG) measurements and analysis of workspace utilization.

Main Results:

• The proposed weight compensation significantly reduced subject effort by an average of 26%.
• The system enabled rehabilitation training throughout the entire workspace of the robotic system.
• EMG measurements confirmed the reduction in muscular effort required by the subjects.

Conclusions:

• Model-based feedforward weight compensation offers an effective solution for enhancing upper limb rehabilitation in stroke patients.
• The developed method is scalable, adjustable, and applicable to various actuated exoskeleton systems.
• This approach facilitates more comprehensive and less strenuous task-related training for individuals with severe motor impairments after stroke.