Exercise Support System Using Audience Robot Behavior

Abstract

We present a VR-based exercise support system featuring virtual spectator robots that adapt their non-verbal behaviors to users’ performance and exercise difficulty, enhancing motivation through dynamic audience effects. While physical activity is crucial for preventing lifestyle-related diseases, sustaining exercise motivation remains challenging. Our system leverages the psychological benefits of audience presence by creating virtual spectators that respond to users during VR boxercise sessions with contextually appropriate cheers, movements, and facial expressions. Unlike static virtual audiences, our robots dynamically adjust their behavioral expressions based on real-time performance metrics and exercise intensity. Experimental evaluation demonstrates that adaptive spectator behaviors significantly enhance users’ enjoyment, sense of accomplishment, and exercise motivation compared to baseline conditions. This work contributes to the growing field of socially assistive robotics by demonstrating how virtual social presence can be computationally modeled to enhance human motivation.