Step on it: asymmetric gain functions improve starting and stopping in virtual reality walking

Abstract

AbstractTransfer functions with a high translational gain can increase the range of walking in virtual reality. These functions determine how much virtual movements are amplified compared to the corresponding physical movements. However, it is unclear how the design of these functions influences the user’s gait and experience when walking with high gain values. In a mixed-methods study with 20 users, we find that their best transfer functions are nonlinear and asymmetrical for starting and stopping. We use an optimization approach to determine individually optimized functions that are significantly better than a common approach of using a constant gain. Based on interviews, we also discuss what qualities of walking matter to users and how these vary across different functions. Our work shows that it is possible to create high-gain walking techniques that offer dramatically increased range of motion and speed but still feel like normal walking.