In October of 2024, the communications satellite Intelsat 33e suddenly lost power and broke into approximately 20 pieces. The satellite, which was launched to deliver communications services to Europe, Africa, and the Asia Pacific region, was positioned in Geosynchronous Orbit (GEO), making its breakup potentially even more unfortunate and impactful.
According to Audrey Schaffer, the Vice President of Strategy and Policy at Slingshot Aerospace, “GEO is accumulating long-lived debris from defunct spacecraft, satellite breakups, and mission-related objects. Unlike LEO, where atmospheric drag helps clear debris over time, GEO lacks the same natural cleansing mechanisms, meaning debris can remain in orbit indefinitely.”
While 20 additional pieces of space debris may not sound like that large of a problem, those 20 pieces of Intelsat 33e joined the more than 25,000 objects larger than 10 cm that the NASA Orbital Debris Program Office claims to exist. And, as Schaffer explained, each of those 25,000 pieces of debris is potentially harmful, “Even small debris fragments can pose a significant threat to operational satellites, which are critical for our economy, national security, and daily lives.”
But it’s not just space debris that can cause problems for satellites in GEO. There is also the challenge of rising congestion in space.
While orders for large GEO satellites are being outnumbered by orders for smaller NGSO satellites among commercial satellite manufacturers, there are still numerous use cases for commercial and military satellites in GEO. In fact, according to the UCS Database, there are almost 600 active satellites in GEO today. That number could grow as more satellites designed to conduct Rendezvous and Proximity Operations (RPO) – such as refueling, transport, and on-orbit servicing – are built and launched.
This is particularly troubling due to the size and cost of the satellites often launched into GEO. As Clayton Swope, the Deputy Director of the Aerospace Security Project and a Senior Fellow in the Defense and Security Department at the Center for Strategic and International Studies (CSIS), explained, “When we’re building satellites for GEO, we’re usually building bigger satellites that are going to be up there for a long period of time. [Satellite operators] put a lot of money into these systems…potentially hundreds of millions of dollars.”
Getting the whole picture
With so many objects that can damage these expensive satellites in GEO, it’s essential that satellite owners and operators know and understand what’s happening in the orbit, where objects are, and how they’re behaving. Unfortunately, GEO is an orbit in which it has historically been challenging to track the movements and behaviors of spacecraft and other objects.
“When it comes to space situational awareness (SSA) for GEO, you just don’t have as many tools in your tool kit. You can’t track objects with radar and see what’s going on,” said Swope. “You’re looking at these objects through telescopes or tracking their RF signals. However, a piece of debris is probably not going to be emitting any signals, which can make identifying and tracking debris even more difficult.”
Worse, the congestion in GEO not only makes it difficult to identify threats that could accidentally damage spacecraft through inadvertent collisions. It also can make it more difficult to identify malicious activities intended to damage satellites.
“Congestion amplifies the challenge of distinguishing between routine satellite maneuvers, potential threats, and anomalous activities. Further, there is potential to take advantage of that complexity to hide assets that might otherwise be routinely monitored,” said Schaffer. “Adversaries could exploit debris fields as ‘natural cover’ to hide their operations, such as by positioning small, maneuverable spacecraft near or within debris clusters to evade detection.”
But it’s not just satellites that have been overlooked that could be a threat. The U.S. government and the Department of Defense (DoD) track satellites launched by adversaries and are often knowledgeable about their location and purpose. But just because a satellite initially behaves in a benign fashion doesn’t mean it’ll stay that way.
“We may think we know what [an adversary’s] satellite is meant to do. For example, we may identify that it’s a refueling satellite. But that satellite probably has the ability to latch onto another satellite, which could make it a threat,” said Swope. “While we might think we know its capabilities, we don’t necessarily know [the owner’s] intentions. One day, that satellite is benign. The next day, there’s a conflict, and that satellite’s now a weapon that can latch onto another satellite and move it.”
With so many threats and the potential for harm to satellites, getting a reliable and accurate SSA and SDA picture is mandatory – even in an orbit where awareness has been challenging. Thankfully, there are new technologies that could help.
Tackling SSA with AI
There are few industries that have been untouched by the emergence of advanced artificial intelligence (AI) and machine learning (ML) technologies. The ability of these technologies to identify things that humans might not see or notice has led to their adoption in practically every field – from healthcare to government to financial services. Now, AI could be the key to identifying threats in space that humans simply can’t see.
“AI and ML can process vast amounts of tracking data in real-time, improving detection, characterization, and prediction of object behavior,” said Schaffer. “ML models can help differentiate between normal and anomalous satellite movements, flagging suspicious patterns that might indicate adversarial activity, and AI can help explain why these suspicious patterns are happening. By enhancing threat detection and predictive analytics, AI-driven SDA solutions enable faster and more accurate decision-making, ultimately strengthening our ability to mitigate threats.”
Not only could AI and ML help satellite owners and operators identify threats more effectively, but they can also help them identify things that aren’t a threat. That can be important when resources are limited, and only so many threats can be given attention at any given time. It can also keep global governments and militaries from overreacting to perceived threats that could ultimately be harmless.
“There may always be ambiguity in GEO, but the goal should be to reduce that. They need to focus their limited resources on what is the actual threat. They need to focus on the satellite that is an actual weapon, not simply a satellite with a camera trying to take pictures of their spacecraft,” said Swope. “They may not want that, but it should elicit a different response. You’re going to respond differently to someone who shows up on your porch with a shotgun than you would if that person has a camera. Space should be no different.”
Numerous companies are now working to build AI SSA and SDA solutions that can help determine what is and isn’t a threat. One of those companies is Slingshot Aerospace, which has developed an AI solution called Agatha, which the company developed in partnership with DARPA.
“Agatha is designed precisely to identify suspicious activity. It evaluates behaviors of objects across time, derives the operational ‘policies’ or intentions of objects by observing their behaviors, and assesses which objects appear to be acting inconsistent with their stated purpose,” explained Schaffer. “Agatha is designed to flag these objects early on so that a military operator receives indications and warnings long before a human analyst might detect anomalies.”
Other companies could soon be joining Slingshot in offering AI solutions for satellites. The startup company, Quindar, recently received $6 million in funding to further develop the AI and ML capabilities of its command and control software – a move that could potentially enable the automation of satellite management. Anduril, a provider of autonomous solutions to the defense industry, is also looking to leverage its advanced AI and ML technologies in space – utilizing AI for both SDA and command and control capabilities.
The earlier warnings these AI solutions enable could make a massive difference for satellite owners and operators looking to protect their spacecraft from threats in GEO. However, AI could have an even more significant impact on the safety of satellites in the future. These new technologies could help ensure that satellite owners and operators make the best choices to ensure the safety of their satellites.
“AI could help operators develop and evaluate courses of action to determine how to respond to attacks,” said Schaffer. “AI can design response options and assess how effective a given option may be to inform operational planners and decision-makers.”
As this technology continues to grow and evolve, even the mere presence and involvement of humans could become unnecessary. As Swope explained, “In the future, will we be able to use AI not just for threat detection but also to mitigate threats in an automated fashion? I’d say yes. If space debris was coming towards a satellite with the possibility of a collision, I could envision AI maneuvering the satellite in an automated fashion, eliminating the need for humans to intervene.”
Explore More:
Data Fusion Leverages All Datasets for Better SDA
DoD’s Increasing Demand for Data Fusion in Space
Non-Earth Imaging (NEI) Brings a Clearer Picture to SDA
