Tempo Synchronization of Physical Activities with a Mixed-Reality Human-Machine-Interface
Sebastian Chinchilla, José Victorio Salazar Luces, Yasuhisa Hirata · 2021 IEEE International Conference on Robotics and Biomimetics (ROBIO) · 2021
Physical trainers aid trainees to increase their motor learning of physical activities by providing multimodal feedback as well as adapting the training to their needs and progress. However, due to the fact that trainers are a scarce resource, added to the social distancing caused by the COVID19 pandemic, not many people have access to a trainer. In order to encourage people to continue practicing physical activities, this work presents a multimodal human-machine interface that automatically coaches ballroom dance. This is achieved by providing haptic, auditory, and visual feedback of the reference position and velocity trajectories, as well as its tempo. The proposed approach was tested in an experiment with 20 subjects. It was found out that subjects were able to reduce their position, velocity, and synchronization errors during and after the training, suggesting that users can retain the learned skills. Furthermore, we observed that higher performance is obtained at faster tempos.