Qualitative Spatial Structure in Complex Areal Objects Using Location-Free, Mobile Geosensor Networks

This paper proposes an in-network, decentralized algorithm for determining the qualitative spatial structure of complex a real objects, which may have multiple disconnected components. The algorithm determines both the containment relations between region components and the adjacency relations between Voronoi regions of those components. Combining this information yields finer-grained qualitative structure of the region, specifically where a component is surrounded (rather than strictly contained) by other region components. Unlike previous work, the algorithm is able to operate in networks of mobile geosensor nodes with no access to coordinate information. The algorithm is evaluated in terms of scalability (communication complexity) and veracity, with the finding that it performs efficiently and correctly.