Snapshots for semantic maps

Abstract

A significant area of research in mobile robotics is in the local representation of a remote environment. In order to include a human in a mobile robot task it becomes important to present the remote information efficiently to a human. A relatively new approach to information presentation is semantic maps. Semantic maps provide more detail about an environment than typical maps because they are augmented by icons or symbols that provide meaning for places or objects of interest. In this paper we present snapshot technology as a means to take pictures from the real world and store them in a semantic map. To make the snapshots and semantic map available to an operator, we identify and discuss general attributes for useful displays and present a mixed reality 3D interface that meets the requirements. The interface and snapshot technology are validated through experiments in real and simulated environments.