Trajectory reconstruction

In this phase, a stream of raw positioning data about moving objects is processed, to obtain ready-to-use trajectories, building a privacy-aware trajectories warehouse.
Reconstruction of trajectories is a challenging problem. The reconstruction accuracy of trajectories, as well as their level of spatio-temporal granularity, depends on the quality of the log entries, since the precision of the position may range from the granularity of a cell of varying size, to the relative (approximated) position within a cell.
Indeed, each moving object trajectory is typically represented as a set of localization points of the tracked device, called sampling. This representation has intrinsic imperfections due to the error measurement of the tracking device and the sampling rate.

This process is dependent on the privacy level which must be enforced, ranging from full identified source data to identity obfuscation. In this context, the privacy issues in traditional databases, which have been actively studied in the recent years, must be extended to cope with new peculiarities arising from trajectories data.