KMI and the SDA TAP Lab

Sheena L. Winder, SDA Analyst

5 minute read

Data lakes describe a paradigm where an enormous amount of data is collected and stored, so that later the data can be organized for easy analysis. Sometimes, the act of organizing and correlating it is left unfinished for too long, and the data lake becomes a data swamp. KMI has joined the Space Domain Awareness Tools, Applications, and Processing (SDA TAP) Lab to bring Department of Defense (DoD) information processing and assimilation up to the speed of relevance.

KMI and the US Space Force share a common goal: ensuring free access to and use of space for all. To meet that mission, debris-causing events must be predicted and avoided. Currently, the DoD provides collision avoidance warning messages for free to all satellite owners and operators. Unfortunately, this service has not prevented a tremendous amount of debris from accumulating in Earth’s orbit. The Iridium-Cosmos collision produced over 2300 pieces of debris, in addition to a vast amount of fragments too small to track. Another significant source of on-orbit debris is the anti-satellite tests of China, India, and Russia which aimed for and struck already orbiting satellites and payloads. Mitigating future debris-causing events requires monitoring both new launches and existing resident space objects (RSOs). However, the US military has neither the equipment nor the personnel to do this on its own. In recent years, the DoD has made great strides in increasing commercial partnerships to build up global sensor coverage using both space and ground infrastructure. Yet, all of these partnerships come at a cost. There is now a deluge of information from different classification levels, in different formats, with different levels of confidence, and of course, all uncorrelated. While the additional information gathered is beneficial, it puts a strain on already stretched DoD resources and personnel. This leaves intelligence analysts and operators from multiple organizations with a plethora of information to process and mesh together, with much of the analysis done manually. This can take hours, days, or even weeks to process, share, deconflict, and coordinate. By that time, it could be too late.

Imagine a scenario where a launch occurs; national, commercial, and private eyes are all watching. Unexpectedly, the launch vehicle goes off track. If it belongs to the US, can the range safety officer still terminate the flight? If it is a foreign launch, there is a scramble to predict where it is going. Whether re-entry, LEO, or GEO, determining the end effect of the failed launch is key to predicting a collision possibility and ensuring necessary avoidance actions can be taken. It's not like the movies. Moving a satellite burns fuel and is a huge decision that shortens the lifespan of a satellite. The information and analysis for the confidence needed to justify moving a satellite might come too late to save it. The operations to command the satellite to move and the actual movement of the satellite will not be instantaneous. Those necessary times are in short supply if a rocket is rapidly approaching the satellite’s position.

There are other examples of space assets that are at risk of flying through a fog of war due to a lack of data. If the tracking of an object is lost due to an extended lack of observations or an unexpected movement, trajectory prediction capability is lost. While valuable time and resources are expended to search for the lost object, more information is continuously rolling in, compounding the tracking and prediction timeline problem. This increases the difficulties for operators and decision-makers. Unexpected changes in position could be the start of an intent to interfere with the operations of a competitor. Another challenging facet is the DoD can only provide collision predictions for objects it is currently tracking.           

Enter the SDA TAP Lab, a tech accelerator host for industry, academia, and government groups to collaborate on a fully automated, commercial SDA system for space sector threat prediction. Over 100 companies and academic institutions have come together to develop this system to rapidly predict, detect, track, identify, and warn in order to protect US military, commercial entities, allies, and partner space systems. SDA contributions range from commercial image scraping and correlating to interception prediction to action recommendation.

KMI’s contribution to the effort consists of analyzing whether an unknown object is actually a known RSO that has moved during a non-tracking period. The data lake is filled with uncorrelated tracks (UCTs). UCTs are usually comprised of multiple observations of a single object which has not been associated with a specific RSO. This yields an unknown object in space that must be assessed and identified as quickly as possible. One avenue of identification is to visually examine the object. But this takes optical resources that may not be available due to location or other priorities. Another avenue is to analyze the characteristics of the signals to determine if there is a match to a known RSO. As mentioned previously, much of this analysis is manual, time-consuming, and lengthy. Obtaining this correlation is vital to space domain awareness— the faster, the better. In less than a minute, KMI’s model produces a list of known RSOs that have a high feasibility of being the unknown object. Through the use of astrodynamical filters, the model compares the unknown object to the most recent orbit of all known RSOs, which is approximately 47,000 objects being tracked by Space-Track. Additional information for timely assessment is produced for the unknown object, such as the probabilities of useful characteristics such as origin, payload, and operability. The resulting information is automatically passed down the pipeline for further processing.

With its membership in the SDA TAP Lab, KMI has provided analysts with a tool that can swiftly and objectively whittle the potential list of candidates down to dozens of higher-probability suspects, hyperfocusing analytical expertise and other intelligence resources. Hours, perhaps days, will be saved in a decision-maker's decision cycle. Whether RSO tracking was lost due to a prolonged observation gap or the RSO performed a concealed or unexpected maneuver, KMI rapidly provides vital information to update understanding of activities in the space domain. Satellite owners and operators are now better equipped to keep freedom of operations available to all.

Recommended column to read next: The First Space Travelers