A Framework for User Adaptation and Profiling for Social Robotics in Rehabilitation

Area of Science:

• Robotics in Medicine
• Pediatric Rehabilitation
• Human-Computer Interaction

Background:

• Pediatric physical rehabilitation faces challenges in patient motivation and adherence.
• Social humanoid robots offer potential for increased engagement and therapeutic assistance.
• Effective rehabilitation requires detailed patient monitoring and personalized feedback.

Purpose of the Study:

• To upgrade the NAOTherapist system for enhanced pediatric rehabilitation.
• To integrate a monitoring system for generating user profile models.
• To improve therapy adaptation and clinical reporting.

Main Methods:

• Development of a robotic architecture (NAOTherapist) for social robot-assisted rehabilitation.
• Implementation of a patient monitoring system to create user profiles.
• Integration of a machine learning algorithm for pose recognition.
• Addition of a clinical report generation system using the QUEST metric.

Main Results:

• The NAOTherapist system now generates user profile models through patient interaction.
• User-adapted therapies are performed based on generated patient profiles.
• Machine learning enables accurate recognition of patient poses during therapy.
• Clinical reports are generated using the QUEST metric for standardized assessment.

Conclusions:

• The upgraded NAOTherapist system effectively supports pediatric physical rehabilitation.
• User profile models and adapted therapies enhance patient engagement and treatment outcomes.
• The integration of machine learning and standardized reporting improves the system's clinical utility.