Duplicated Virtual Co-embodiment Displaying Out-of-view Movements for Tennis Training

Motor skill instruction in virtual environments allows for the presentation of different body information from physical environments, thereby supporting more efficient learning. Virtual Co-embodiment, in which a learner and teacher share control of a single avatar, has demonstrated promising results in motor skill acquisition. However, existing methods are typically limited to movements within the user’s field of view, making their applicability to sports involving out-of-view movements uncertain. To address this issue, we developed a system that displays duplicated avatars in front of the user, allowing the user to observe and adjust their movements more easily. In this study, we focused on tennis return shots and displayed the teacher’s, the learner’s, and the co-embodied avatars, enabling users to recognize and correct differences in their movements through virtual co-embodiment. We conducted a user study on hitting back a tennis ball. We found that providing learners with too much information at once makes correcting their movements difficult. The comments also suggested that reducing the amount of information shown to learners or slowing down their movements may be necessary.