Ubitrack: Automatic Configuration of Pervasive Sensor Networks for Augmented Reality
Pervasive Computing, IEEE, PP (99):
1 (2010)DOI: 10.1109/MPRV.2010.50
Abstract
The range of most Virtual Reality and Augmented Reality systems is usually constrained to laboratories by the limited range of a single expensive 6DOF tracking system. Our Ubiquitous Tracking (Ubitrack) approach seeks to overcome this limitation, not by extending the range of a single tracker, but by dynamically incorporating a number of heterogeneous sensors. The Ubitrack tool set consists of Spatial Relationship Graphs (SRGs), which allow for a formal specification of tracking environments, and Spatial Relationship Patterns, which are used to derive data flow configurations by systematic analysis of the SRG. This formal framework supports a distributed software architecture, where clients efficiently produce, transform, transmit and consume tracking data, coordinated by a Ubitrack server. In this document, we discuss both theoretical and implementational aspects and illustrate the new possibilities on a scenario, where a mobile AR user is guided through a building using a diverse and changing set of sensors.