Forensic Tracking and Mobility Prediction in Vehicular Networks

Abstract

Most contemporary tracking applications consider an online approach where the target is being tracked in real time. In criminal investigations, however, it is common that only offline tracking is possible, where tracking takes place after the fact; that is, given an incomplete trace of a suspect, the task is to reconstruct the missing parts and obtain the full trace. With the recent proliferation of modern transportation systems, target entities are likely to interact with different transportation means. Thus, in this paper, we first introduce a class of mobility models that has been especially tailored for forensic analysis then propose several instances emulating different transportation means. We then use these models to build a full-fledged offline multi-modal forensic tracking system that reconstructs an incomplete trace of a particular target. We provide theoretical evaluation of the reconstruction algorithm and show that it is both complete and optimal.