System uses AR feedback to train athletes – or to help rehab patients
Whether you're an athlete looking to improve your performance, or a rehabilitative therapy patient learning to walk again, it helps if you can be shown what to do differently as you're in motion. Well, that's exactly what a new augmented reality-based (AR) system is designed to do.
The technology is being developed at Canada's University of Alberta, by a team led by mechanical engineering Asst. Prof. Hossein Rouhani. Assisting in the research is Michael Xiu, an undergrad student from China's Sun Yat-sen University – he's part of the Mitacs Globalink summer internship program, in which international students get to help solve innovation challenges at Canadian universities.
Users of the U Alberta system are outfitted with small retroreflective markers on key parts of their bodies, which are precisely tracked by an array of infrared cameras as they run, walk, skate or otherwise move about. Those markers are combined with inertial measurement units (IMUs), which also track the user's movements via a combination of accelerometers and gyroscopes.
Plans call for the system to ultimately be IMU-only, so no cameras will be required. This means that it could be used by athletes in the field, where they would be performing in a natural manner.
Motion data captured by the system is transmitted to a computer, which uses custom algorithms to build a real-time animated wire-frame model of the user. That 3D model is compared to one that was obtained from someone who was performing the same activity the "right" way.
Thanks to software being developed by Xiu, users will soon see both models overlaid on their view of the real-world environment, via a pair of AR goggles. Coaches or therapists will additionally see both models on the computer screen.
"[The user] can see their own motion plus the targeted motion that is either based on an able-bodied individual in the case of patients, or based on the motion patterns of advanced-level professional athletes," Rouhani tells us. "They can use the difference between their actual motion and the targeted motion as feedback, and that feedback-based learning would expedite the training outcome."
A prototype version of the system should be in use for evaluation of the university's sports teams, within the coming year.