MoCapGym: Erlernen motorischer Fähigkeiten mittels Motion-Capture

Performing a precise sequence of movements is essential in sports and physical therapy. These exercises have to be repeated regularly, frequently and in a correct way to improve the skills. In most training sessions, neither a teacher nor a therapist can be present in order to ensure a correctly performed exercise. Current tools for learning motion skills autonomously are based on training videos, animations or textual description without the opportunity for receiving feedback. Performing an exercise incorrectly can therefore lead to limited success or even cause negative consequences. In this work we developed the novel Virtual Reality application MoCapGym, which allows users to interactively and autonomously learn and improve complex motions. Based on an evaluation of different approved learning theories and concepts in the area of motion learning, a novel multi-phase learning concept for 3D virtual environments was developed. The approach combines traditional aspects as well as tries to exploit the full potential of VR devices for motion capture and 3D visualization. With the help of an optical tracking system and an active motion capture suit, it was possible to design a learning environment for arbitrary and complex motions. An automatic visual feedback concept correlates real time motion data with recorded optimal exercises. Discrepancies are visualized in a 3D environment. A user study shows that the developed learning environment can be used for learning body movements in a short time frame.