Inferring Position Knowledge from Location Predicates

Many context- and location-aware applications request high accuracy and availability of positioning systems. In reality however, knowledge about the current position may be incomplete or inaccurate as a result of, e.g., limited coverage. Often, position data is thus merged from a set of systems, each contributing a piece of position knowledge. Traditional sensor fusion approaches such as Kalman or Particle filters have certain demands concerning the statistical distribution and relation between position and sensor output. Negated position statements ("I'm not at home"), cell-based information or external spatial data are difficult to incorporate into existing mechanisms. In this paper, we introduce a new approach to deal with different types of position data which typically appear in context- or location-aware application scenarios.

• Type: Conference
• Year: 2007
• Where: LOCA
• Authors: Jörg Roth
• Keywords: Current Position, LOCA 2007, Location-aware Application, Position Data