Uploaded June 2020 | Updated September 2026, 3 weeks ago
Developed by Sandia National Labs, Tracktable is a set of trajectory analytics used to apply advanced machine learning techniques to large trajectory data sets. It enables searches for shapes and patterns in space and time by providing a mathematical framework to describe such patterns. Given that framework, Tracktable provides tools for fast search and categorizations in order to organize, search, and quickly analyze millions of patterns enabling the computer to group similar shapes together and to find unique or unusual trajectories without first requiring humans to define “normal”.
Additionally, its ability to treat time as a variable, similar to space, enables searches for collective behavior and patterns over long periods of time.
Finally, fast search techniques enable Tracktable to predict the paths and destinations of moving objects by comparing their observed paths to historical databases of trajectories.
The notion of fundamentally representing the trajectory as a vector of “features” is the key idea that makes this approach different from previous trajectory analysis techniques.
More information is available at tracktable.sandia.gov/.
SAND2020-5778V
Developed by Sandia National Labs, Tracktable is a set of trajectory analytics used to apply advanced machine learning techniques to large trajectory data sets. It enables searches for shapes and patterns in space and time by providing a mathematical framework to describe such patterns. Given that framework, Tracktable provides tools for fast search and categorizations in order to organize, search, and quickly analyze millions of patterns enabling the computer to group similar shapes together and to find unique or unusual trajectories without first requiring humans to define “normal”.
Additionally, its ability to treat time as a variable, similar to space, enables searches for collective behavior and patterns over long periods of time.
Finally, fast search techniques enable Tracktable to predict the paths and destinations of moving objects by comparing their observed paths to historical databases of trajectories.
The notion of fundamentally representing the trajectory as a vector of “features” is the key idea that makes this approach different from previous trajectory analysis techniques.
More information is available at tracktable.sandia.gov/.
SAND2020-5778V
