High-performance spatiotemporal trajectory matching across heterogeneous data sources

Gong, Xuri, Zhou Huang, Yaoli Wang, Lun Wu, and Yu Liu. “High-Performance Spatiotemporal Trajectory Matching across Heterogeneous Data Sources.” Future Generation Computer Systems 105 (November 1, 2019). https://doi.org/10.1016/j.future.2019.11.027.

A few months after this paper had been published I was confronted with a similar problem. At the time I wanted to automatically figure out which pilots were flying in which aircraft. This paper attempts to answer the question how one can figure out which objects are moving through spacetime together, or suspiciously close to one another.

It is this paper which presents a brilliantly simple algorithm for figuring that out, which can also be applied at scale. Although incredibly powerful, it should be noted that this sort of technology is also incredibly useful for surveillance. Stuff like this is among the reasons you do not take your phone with you to protests. It’s able to combine all sorts of different data sources, figure out what devices belong to you, who you have been with, where, when and for how long. The only way to prevent this information from being collected from you is by ensuring this information is not collected in the first place, or alternatively to mislead the systems.

See these pages as well for more background to an actual implementation of this paper as well.