Robot-Mediated gesture-based memory game for older adult psychophysical stimulation

The rapid growth of the aging population in developed countries makes healthcare an important social challenge. In this context, service robots can play a key role. This work presents a software application for a service robot (TIAGo, PAL Robotics) implementing a motor-cognitive game. The activity combines cognitive and physical stimulation, a design that is relatively uncommon in literature. The embodied interaction design of the game distinguishes it from classical touchscreen-based games. In the game, the robot mimes letters with its arm, and the user has to recognize and then imitate them. The letter sequence increases in length each turn to train memory. User gestures are tracked using an ArUCo marker, and classified via a neural network. The application was tested on 10 young subjects and 4 community-dwelling older adults (82.3 ± 3.5 years). Recognition accuracy reached 92.2% and 80.5%, respectively, for young and older adults. Post-Session questionnaires highlighted high engagement and perceived usefulness, especially among older users who appreciated the memory and physical training aspects. This pilot project demonstrates the potential of integrating service robots into eldercare to support both patients and caregivers.