Score-based recommendation for efficiently selecting individual virtual agents in multi-agent systems
PubDate: November 2017
Teams: RWTH Aachen University；King Mongkut’s University of Technology
Writers: Andrea Bönsch；Robert Trisnadi；Jonathan Wendt；Tom Vierjahn；Torsten W. Kuhlen
Abstract
Controlling user-agent-interactions by means of an external operator includes selecting the virtual interaction partners fast and faultlessly. However, especially in immersive scenes with a large number of potential partners, this task is non-trivial.
Thus, we present a score-based recommendation system supporting an operator in the selection task. Agents are recommended as potential partners based on two parameters: the user’s distance to the agents and the user’s gazing direction. An additional graphical user interface (GUI) provides elements for configuring the system and for applying actions to those agents which the operator has confirmed as interaction partners.
