Immersive augmented reality system for the training of pattern classification control with a myoelectric prosthesis

Area of Science:

• Rehabilitation Engineering
• Human-Computer Interaction
• Prosthetics and Orthotics

Background:

• Hand amputation significantly impacts quality of life.
• Myoelectric prostheses offer restored motor function but pose training challenges.
• Virtual and augmented reality (AR) can enhance prosthesis training engagement.

Purpose of the Study:

• To present a novel AR training framework for myoelectric prosthesis control.
• To evaluate the framework's effectiveness in skill transfer from simulated to real tasks.
• To assess user experience, including motivation and perceived difficulty.

Main Methods:

• Developed an AR framework integrating virtual elements into a real-world, first-person view.
• Evaluated the framework with 13 able-bodied subjects and 1 limb-deficient person.
• Intervention group trained with AR (simulated clothespin task with visual feedback); control group did not.

Main Results:

• The intervention group showed significant improvement in the posttest compared to the control group.
• Intervention group transferred more pins (28 vs. 14.5) and dropped/broke fewer pins (1/5 vs. 3.5/14.5).
• Intervention group reported lower difficulty (5.2 vs. 7.1), higher fun (8.7), and better feedback utility (8.5).

Conclusions:

• The AR system effectively transfers skills from simulated to real tasks.
• The framework provides a positive and motivating user experience.
• The developed AR system is open-source for further development.