Approximate order-k Voronoi cells over positional streams

  • Authors:
  • Kostas Patroumpas;Theofanis Minogiannis;Timos Sellis

  • Affiliations:
  • National Technical University of Athens, Hellas;National Technical University of Athens, Hellas;National Technical University of Athens, Hellas

  • Venue:
  • Proceedings of the 15th annual ACM international symposium on Advances in geographic information systems
  • Year:
  • 2007

Quantified Score

Hi-index 0.00

Visualization

Abstract

Handling streams of positional updates from numerous moving objects has become a challenging task for many monitoring applications. Several algorithms have been recently proposed for providing exact answers particularly to continuous range and k-nearest neighbor queries against current object positions. In this work, we introduce a processing technique for efficiently maintaining an approximate order-k Voronoi cell around a certain point of interest when all objects continuously change their locations. This heuristic can easily provide a fairly reliable estimate of the k-nearest neighbors for any query point found inside the constructed cell. We further extend our method to handle positional updates that are not received concurrently for all objects, but instead remain valid for a specific time interval according to a sliding window model. Extensive experimental analysis over synthetic datasets confirms the robustness and scalability of this approach offering near real-time cell maintenance with acceptable error margins.