Sketch on Lifelong AR Agents in U-VR Environments

In this paper, we present a lifelong AR agent that provides the customized assistance with a mobile user at anywhere, anytime, anything. The agent infers the user's characteristics, e.g., ability, knowledge, and preference, based on the history of the user's context in U-VR environments. It differentiates the assistance generation by reflecting the user's characteristics. The agent also shows appropriate responses to environmental conditions where the user actually exists. To show possibilities of our agent, we have developed applicable applications how we could apply it to smart home environments. Therefore, we expect the potential for the proposed lifelong learning AR agent to be used as a personalized assistant in U-VR environments.