Preserving privacy of moving objects via temporal clustering of spatio-temporal data streams

  • Authors:
  • Roland Assam;Thomas Seidl

  • Affiliations:
  • RWTH Aachen University, Germany;RWTH Aachen University, Germany

  • Venue:
  • Proceedings of the 4th ACM SIGSPATIAL International Workshop on Security and Privacy in GIS and LBS
  • Year:
  • 2011

Quantified Score

Hi-index 0.00

Visualization

Abstract

The collection and exploration of a profusion of location data for mobility pattern mining has been met with stiff resistance from individuals and privacy groups due to privacy and security concerns. Such concerns have driven research to find the best trade-off between data mining and privacy. Unlike other techniques that make use of the spatial properties of spatio-temporal data to attain k-Anonymity, this study is inclined at exploring the possibilities of enforcing k-Anonymity by adroitly using the temporal domain of a spatio-temporal data. We propose a novel technique to achieve k-Anonymity for moving objects through temporal clustering. Our approach first perturbs and delays a trajectory trace once it arrives at a trusted server. This delay translates into the periodic accumulation of traces. Based on a system specified threshold distance and some newly defined time notions, redundant delayed traces are pruned and then similar traces are clustered. Traces in the same clusters are assigned the same spatial and time value, thus making them quasi-identifiers of the k-Anonymity privacy paradigm. We illustrate through empirical experiments the effectiveness of our technique on a real and synthetic dataset; and show that our approach provides strong privacy through k-Anonymity.