Virtual Reality-Based Gymnastics Visualization Using Real-Time Motion Capture Suit

Gymnastics imposes great physical and mental stress on its athletes. Whether due to injury, fatigue, or a desire to rest before a competition, gymnasts cannot always practice their skills successfully and safely in training. To account for this, gymnasts often close their eyes and imagine themselves performing their moves to train focus and muscle memory without excessive physical exertion. This paper presents a system that allows gymnasts to take this a step further by controlling a virtual gymnast from the first-person perspective in virtual reality (VR). When operating the system, the user wears a full-body motion capture suit and a VR headset. The user controls a virtual gymnast using gestures tracked by the suit and perceives an immersive virtual environment through the VR headset. To ensure a fruitful user experience, the virtual environment must be immersive and realistic. First, the virtual gymnastics movements must mimic those of real life. Second, the system must have a high image- rendering speed, dropping no lower than 60 frames per second. Third, real-life movements must correspond to virtual movements nearly instantly, so the motion-to-display (MTD) latency should be minimized. In this paper, we demonstrate that the system's virtual gymnastics animations are derived directly from moves performed in real life using motion capture suit. Additionally, the system achieves an average frame rate of 71.46 frames per second and an average MTD latency of 59.31 ms.