We present a pilot study comparing visual feedback mechanisms for preventing physical collisions between co-located VR users. These include Avatar (a 3D avatar in co-located with the other user), BoundingBox (similar to HTC's "chaperone"), and CameraOverlay (live video feed overlaid on the virtual environment). Using a simulated second user, we found that CameraOverlay and Avatar had the fastest travel time around an obstacle, but BoundingBox had the fewest collisions at 0.07 collision/trial versus 0.2 collisions/trial for Avatar and 0.4 collisions/trial for CameraOverlay. However, subjective participant impressions strongly favoured Avatar and CameraOverlay over BoundingBox. Based on these results, we propose future studies on hybrid methods combining the best aspects of Avatar (speed, user preference) and BoundingBox (safety).

Additional Metadata
Keywords Co-location, Collision avoidance, Multi-user, VR
Persistent URL dx.doi.org/10.1145/3027063.3053180
Conference 2017 ACM SIGCHI Conference on Human Factors in Computing Systems, CHI EA 2017
Citation
Scavarelli, A. (Anthony), & Teather, R. (2017). VR collide! Comparing collision-avoidance methods between colocated virtual reality users. In Conference on Human Factors in Computing Systems - Proceedings (pp. 2915–2921). doi:10.1145/3027063.3053180