Movement recognition and preference in home-based robot-assisted stroke rehabilitation

Robots can be effective tools for rehabilitation of subjects with stroke. Furthermore, home-based robotic rehabilitation could reduce the costs and improve the therapy outcome. We worked on such a context within the SCRIPT (Supervised Care and Rehabilitation Involving Personal Telerobotics) project. We designed a system composed of a wearable passive orthosis which assists and measures hand and wrist movements, a personal computer and motivational and interactive games. In this paper, we focused on the definition of the movements which are used to play the interactive games. We considered the results of testing our methods on 20 subjects with chronic stroke who completed a six weeks clinical trial and investigated their preference of certain movements. Our results show a tendency to train hand movements among subjects with lower impairment and wrist movements for more impaired participants.