Improved Takeover System: Enabling Single Actor Control for Multi-Avatar Populated Virtual Realities
Year: 2024
Type: conference-paper
Abstract: Social virtual reality (SVR) has become more accessible to the general public. Just as in real life, users expect the possibility of engaging with many others within SVR. Creating open-ended, densely populated environments with highly interactive virtual humans remains challenging. Despite the advances in artificial intelligence and language understanding, we still see the need for human intervention to control unanticipated scenarios or answer questions. Previous work has allowed a single actor to take over control of multiple virtual humans when interaction with the user was needed. However, those works focused only on non-verbal interaction. This paper proposes a system that supports full-body avatars with six-point tracking and a simpler take-over procedure. We demonstrate a scenario that enables both verbal and non-verbal interactions.
Authors: Jingyi Zhang, Anthony Steed
Institution: University College London
Language: English
Subfield: Human-Computer Interaction
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W4404915902