As the number of deployed satellites continues to grow at stunning rates, it is becoming increasingly difficult for the military to analyze the deluge of inbound data it receives from its space assets in relevant, decision-making timeframes. By adopting technologies like artificial intelligence (AI) throughout its space architecture, the military can transform how it analyzes its data in ways that can ensure the delivery of critical information to key decision-makers at the speed of conflict before adversaries strike.
Earlier this month at the SATELLITE 2023 conference, artificial intelligence and satellite experts across commercial industry convened during a special panel, “How AI and Space Technologies Combine to Benefit the Critical Mission,” to explore the different applications, benefits and some threats AI can deliver to the U.S. military’s space initiatives.
Space, the military, and AI
One fact that all the panelists agreed on was that artificial intelligence, in general, is a technology meant to extract humans out of routine operational functions. According to SpiderOak & York Space Systems’ Charles Beames, “What it does is it replaces people.”
“Everything we do in space, we do it for the data,” explained Beames. “And a big part of creating data is doing the analytics to make [data] useful. Rather than having thousands of people looking at each piece of data, they can deploy these great algorithms…that can actually be a huge force multiplier.”
Lockheed Martin’s Johnathon Caldwell brought up the point that the relevance of data has a short lifespan, as speed is a dominant factor in the space domain. “With the sensors we have on orbit and with people in the loop, we have a hard time today keeping up with analyzing the data,” explained Caldwell. “The human factor is the limiting factor.”
He explained that as commercial industry and the military build satellite sensors to proliferation, humans on the ground are going to be unable to keep pace with the sheer volume of incoming data. “It’s not data that policymakers and military leaders need, but rather knowledge and information to be able to make decisions,” said Caldwell. “To process the volume of data that’s going to be coming off of the sensors, networks, and systems is going to require us to move into a new era of how we think about looking at that data.”
When reframing how data is regarded, it is critical to remember that data is not always relevant, and that it will not stay relevant forever. While it’s been established that the military and federal government has a problem keeping up with data volumes, they also have a greater issue of sifting through that data – at the speed of conflict – and decide which information is relevant to decision-making.
“We have to clean the table…and get on to the relevant data,” said Caldwell. “It all happens at such an amazing tempo. The speed of space is already high, and the speed of conflict will amp up the timetable in which decisions need to be made. And it’s going to be much quicker than any of us anticipate.”
AI can simplify data complexities
By leveraging AI within their space architectures, the federal government and military can have the ability to analyze information faster and automate some of the more routine – yet extremely complex – processes.
According to SES Space & Defense’s Ram Rao, at the heart of AI are the complexities involved in network systems. “Every system is huge,” said Rao. “For example, SES Space & Defense’s O3b mPOWER satellites are going to be operational by the end of this year, and each of those satellites will have 5,000 beams. With 11 satellites in tow, the O3b mPOWER constellation will, in total, have 55,000 beams. There has to be resource management systems which can really control all those beams and complexities that come with it.”
Rao explained that the amount of incoming and outgoing data that these satellites will be processing cannot be managed by humans alone. Factors like power, bandwidth, and interference management, along with beam switching, hopping, shaping, and formatting, will require more than just traditional conventional algorithms, machine learning to handle vast amount of data as well as deep learning algorithms with neural networks adapting and learning from the data.
“Approaching conventional management methods makes it very difficult to really address the requirements,” explained Rao. “Especially when it comes to the speed of implementing.” He went on to explain that if the military were to execute a mission and needed to switch from one satellite beam to another beam, data computing must occur extremely fast to ensure seamless mission communications.
Especially in times of crisis or conflict, if adversaries were to target U.S. military or government satellites, AI technology could detect attacks before they occur, and switch services over to other satellites in the same orbit, or in a different orbit altogether. By being able to sense and elude an enemy’s jamming, interference or degradation of U.S. space assets, the military would have created a resilient space architecture that is capable of denying any attempts adversaries were to make to interrupt critical missions.
“SES Space & Defense’s specialization is end-to-end connectivity, which includes space, satellites, and ground systems,” said Rao. “If there is a degradation or jamming trend that is occurring on-orbit, AI could alert human operators to the trend and ensure that those kinds of critical issues are addressed. Managing those things and making sure that the satellites and systems are healthy is very important. That can be done, but not just through manual, higher-level monitoring. It has to be at a very low – and very intelligent – level. That is an example of when AI becomes critically important.”