CT In order to support location-aware applications it is necessary to locate people and equipment in near real-time. To avoid unnecessary exposure of details of the underlying tracking and positioning systems, researchers have proposed an additional layer of indirection between sensors and applications: a location service. In this paper, we examine how such a service should acquire and integrate location data from multiple heterogeneous location subsystem. We propose a fusion algorithm based on a formally de ned, hierarchical location model. The algorithm can identify and exploit overlaps among location sightings to improve accuracy. Moreover, inconsistencies can be detected and dealt with either by nding the least common denominator, or the most likely alternative.