This part of my research has been going on for a rather long time, and has addressed various issues
related to the notion of uncertainty in the settings of moving objects trajectories.
The sources of uncertainty in this context are plentiful: from the mere
fact that the positioning devices are inherently imprecise, to the pragmatic aspect
that, although the objects are moving continuously, location-based servers can only
be updated in discrete times. Hence come the problems related to modelling and
representing the uncertainty in Moving Objects Databases (MOD) and, as a consequence,
problems of efficient algorithms for processing various spatio-temporal
queries of interest. Given the ever-presence of uncertainty (since the dawn of philosophy
through modern day nano-level science), systematic approaches are needed to analyze the syntactic constructs as well as the corresponding processing algorithms for various predicates capturing the semantic impacct of uncertainty on spatio-temporal queries.

A flip-side of the uncertainty due to devices/measurements, is the quest to reduce the storage requirements for the large quantities of (location,
time) data in MOD servers, as well as limit the transmisison during the tracking process in distribued enviroments. Clearly, this calls for applying data-reduction techniques, to which part of my work has been dedicated over the years.
Reducing the size of the data-points will ultimately introduce an uncertainty,
and various trade-offs: e.g., the size vs. imprecision vs. energy (in case of communication) need to be managed.


Below are some specific problems and the corresponding sample-publications:
  1. Managing Uncertainty in Moving Objects Databases
  2. Uncertainty of Spatial Trajectories
  3. Spatio-Temporal Data Reduction with Deterministic Error Bounds
  4. Real-time vs. Historic Spatio-Temporal Data Compression
  5. Uncertain Range Queries for Necklaces
  6. Ranking Uncertain Trajectories for NN Queries this is the official version as appeared in the VLDBJ in 2011. The version which was reviewed and accepted had 30 pages which had to be cut down to 25, so some of the experimental observations were not reported. The full-version is available as a Technical Report from the EECS Dept. website (here). The source code implementing our algorithms is available here...
  7. Uncertainty on Road Networks
  8. Fusing Uncertainty from Heterogeneous Sources
  9. Data Compression when tracking in WSNs
  10. More recently, part of my efforts has focused on fixing inconsistencies in MOD, which arise due to modeling the continuous nature of the motion via discrete samples. This kind of model-induced inconsistencies cannot be tackled/handled with the traditional techniques for Integrity Constraints management. The interest in this type of problems was motivated by the many discussions with Roberto Tamassia about a class of snap-rounding and geometric degeneracies removal via perturbation techniques. Subsequently, there were several back-and-forths with the folks from Ludwig Maximilian Univ. this is the paper that appeared in SSTD 2015.
  11. Part of the recent efforts focused on fusing uncertain location data from heterogeneous sources for mobile objects