Adapting Task Difficulty in a Cup-Stacking Rehabilitative Task

As the need for accessible upper-body stroke rehabilitation grows, it becomes increasingly important to investigate how the difficulty level of rehabilitation tasks can be personalized to a patient and automatically adapted based on the patient's progress in therapy. We introduce a framework that uses Fitts' law to define task difficulty and iteratively apply it to dynamically adjust difficulty levels and to assign therapy tasks within the context of a cup-stacking occupational therapy activity. Our preliminary simulation results support the hypothesis that the model can adapt its difficulty levels based on a user's time taken to stack a cup at various points on a table. Future work includes exploring the impact of different variables on the model's adaptability and integrating personalized verbal feedback from a socially assistive robot.