Intelligent Interaction System Capable of Autonomous Mobility for At-Home Workouts

In this paper, we present an intelligent interaction system capable of autonomous mobility for at-home workouts. Whereas the current interaction systems for at-home workouts have had much difficulty in mobility, the presented system is able not only to assist people to work out at-home through the use of advanced deep learning technologies such as action recognition, human pose estimation, and bio-signal recognition but also to have autonomous mobility through the utilization of a mobile robot. To verify the system’s feasibility with regard to intelligent interaction capable of autonomous mobility, we have conducted individual experiments for each module in the system such as mobile robot navigation, interfacing, unity contents visualization, action recognition, human pose estimation, bio-signal recognition, and human data generation.