Related People:

Blair MacIntyre

Cindy Robertson

Associated Projects

Accounting for Uncertainty in Mobile AR Systems

AIBAS: Adaptive Intent-Based Augmentation System

paperpic missing   Adapting to Registration Error in an Intent-Based Augmentation System
An ongoing research problem in Augmented Reality (AR) is to improve tracking and display technology in order to minimize registration errors. However, registration is not always necessary for users to understand the intent of an augmentation, especially in industrial applications where the user and the system have extensive semantic knowledge of the environment. In this chapter, we review the ideas of communicative intent developed for desktop graphical explanation systems by Seligmann and Feiner, and discuss how these approaches are the basis for our hypothesis that semantic knowledge of a scene can be used to ameliorate the effects of registration errors. We describe a set of AR visualization techniques for augmentations that adapt to changing registration errors. We first define a set of strategies that use semantic knowledge of the augmentation to enhance the augmentations with additional contextual cues. These context cues help users understand the intent of the augmentation in the presence of registration error. We then introduce algorithms that use features and feature points on objects to control these strategies in the presence of changing registration errors. Finally, these algorithms and techniques are demonstrated in four maintenance situations that challenge a userís ability to interpret the semantics of a scene.

Full Reference:

C. Robertson and B. MacIntyre, 2003. "Adapting to Registration Error in an Intent-Based Augmentation System," In: Virtual and Augmented Reality Applications in Manufacturing, ed. S.K. Ong and A.Y.C. Nee, London, Springer Verlag, in press.

Related Links

Go to the main Georgia Tech site
Go to Georgia Tech GVU site