Human Upper-Body Inverse Kinematics for Consumer Grade Virtual Reality

In virtual reality (VR), being able to see your own virtual avatar instead of floating head and hands can enhance the immersion. However, most consumer VR devices only provide limited tracker data. In order to reconstruct human upper body movements from rare tracking data, we propose the HeadIK method based on human inverse dynamics optimization, and our research shows that this method can achieve the reconstruction of upper body movements using only the head-mounted display and the handle, which can enhance the immersion of VR experience.