Architecture of an Automated Coaching System for Elderly Population

We present an automated coaching system for elderly population living in assisted homes. The system guides its users through a sequence of exercises and tests. Each exercise is demonstrated by a pre-recorded video of a coach, checked for correct execution and qualitatively evaluated. Automatic coaching advices are generated in order to improve the execution. Performance measurements are shown as an immediate feedback to the user, and stored and evaluated over time. The system is designed to allow for a remote interaction with a coach, and, to bolster social aspect of the exercise, for concurrent exercise of two (or eventually multiple) remote users.

[1]  Andrew W. Fitzgibbon,et al.  Real-time human pose recognition in parts from single depth images , 2011, CVPR 2011.

[2]  J. Krakauer.  Motor learning: its relevance to stroke recovery and neurorehabilitation. , 2006, Current opinion in neurology.

[3]  Justin W. L. Keogh.  Able bodies balance training , 2009 .

[4]  Stepán Obdrzálek,et al.  Accuracy and robustness of Kinect pose estimation in the context of coaching of elderly population , 2012, 2012 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[5]  Heidi Sveistrup,et al.  Motor rehabilitation using virtual reality , 2004, Journal of NeuroEngineering and Rehabilitation.