To enable rich and meaningful Augmented Reality (AR) experiences within a Ubiquitous Computing environment, a detailed, coherent and up-to-date spatial model of the world is essential. However, current tracking technologies are limited in their range and operating environments. This has, so far, restricted the development of wide-area AR applications. To extend the range of AR applications, it will be necessary to combine widely different tracking technologies dynamically, aggregating their data and balancing their trade-offs. In this paper, we propose a formal framework, called Ubiquitous Tracking, which uses a graph-based model of spatial relationships to build dynamically extendible networks of trackers suitable for the high-precision, low-latency requirements of Augmented Reality. The framework is powerful, allowing us to model existing complex tracking setups; extensible, accommodating new trackers, filtering schemes and optimisation criteria; and efficient, allowing an effective implementation within existing AR systems.
Reference currently not available.