Distributed Tracking with Multiple Sensors for Augmented Reality

Augmented Reality (AR) applications traditionally have been restricted to a small working volume, since current high-precision tracking technology is clearly limited in the range of operation. This paper presents ongoing work on creating a highly distributed and dynamic middleware for the aggregation of an arbitrary amount of location sensors. It is designed to work in a peer to peer fashion to allow the integration of mobile setups into stationary ones or vice versa and the ad-hoc connection of two or more mobile setups at runtime. This allows AR to be used as an interfacing technology in Ubiquitous Computing environments. The middleware is based on a formal model of attributed spatial relationship graphs that allow a uniform treatment of typical sensors used for AR. I present the distribution strategy and a generic distributed algorithm for selecting an optimal combination of available sensors, with optimality being defined differently by each application using the middleware. A discussion of intermediate results, observed problems and possible remedies concludes the paper.     «

Augmented Reality (AR) applications traditionally have been restricted to a small working volume, since current high-precision tracking technology is clearly limited in the range of operation. This paper presents ongoing work on creating a highly distributed and dynamic middleware for the aggregation of an arbitrary amount of location sensors. It is designed to work in a peer to peer fashion to allow the integration of mobile setups into stationary ones or vice versa and the ad-hoc connec...     »