Comparing Four Approaches to Generalized Redirected Walking: Simulation and Live User Data

Abstract

Redirected walking algorithms imperceptibly rotate a virtual scene and scale movements to guide users of immersive virtual environment systems away from tracking area boundaries. These distortions ideally permit users to explore large and potentially unbounded virtual worlds while walking naturally through a physically limited space. Estimates of the physical space required to perform effective redirected walking have been based largely on the ability of humans to perceive the distortions introduced by redirected walking and have not examined the impact the overall steering strategy used. This work compares four generalized redirected walking algorithms, including Steer-to-Center, Steer-to-Orbit, Steer-to-Multiple-Targets and Steer-to-Multiple+Center. Two experiments are presented based on simulated navigation as well as live-user navigation carried out in a large immersive virtual environment facility. Simulations were conducted with both synthetic paths and previously-logged user data. Primary comparison metrics include mean and maximum distances from the tracking area center for each algorithm, number of wall contacts, and mean rates of redirection. Results indicated that Steer-to-Center out-performed all other algorithms relative to these metrics. Steer-to-Orbit also performed well in some circumstances.