Modeling Personalized Difficulty of Rehabilitation Exercises Using Causal Trees

Area of Science:

• Robotics
• Rehabilitation Medicine
• Human-Computer Interaction

Background:

• Rehabilitation robots enhance patient motivation through game-like interactions.
• Personalized exercise difficulty is crucial for maximizing rehabilitation outcomes and engagement.
• Existing methods use generic difficulty, failing to account for individual differences in perception.

Purpose of the Study:

• To develop a method for calculating personalized exercise difficulty for stroke survivors using rehabilitation robots.
• To address the limitations of generic difficulty settings in rehabilitation exercises.
• To improve both rehabilitation outcomes and user motivation.

Main Methods:

• Formulated a causal tree-based method to calculate exercise difficulty.
• Utilized user performance data to determine individual difficulty levels.
• Focused on the unique perceptions of exercise difficulty reported by stroke survivors.

Main Results:

• The causal tree-based method accurately models individual exercise difficulty.
• The approach provides an interpretable model of exercise difficulty for users and caretakers.
• Demonstrated that stroke survivors have varied and unique perceptions of exercise difficulty.

Conclusions:

• Personalized difficulty adjustment is key for effective robot-assisted rehabilitation.
• The proposed method offers a more accurate and interpretable approach to difficulty scaling.
• This research has implications for improving stroke survivor recovery and engagement with robotic therapy.