Big Ideas Lab - STARMOC

Episode Date: August 4, 2026

The number of objects in orbit is doubling every two years, and the systems built to keep watch still run on formats designed around 1960s punch cards. At Lawrence Livermore National Laboratory, a mission operations center called STARMOC is proving that the future of space operations can be built fast, affordable and in-house - by testing new ways to fly satellites, train AI and stay ahead in a crowded, contested sky. This is STARMOC at Lawrence Livermore National Laboratory. Guests featured (in order of appearance): Luc Peterson - Associate Program Leader for Data Science, Space Program, LLNL Phillip Rittmuller - Principal Investigator for STARMOC -- Big Ideas Lab is a Mission.org original series. Executive Produced by Levi Hanusch. Sound Design, Music Edit and Mix by Matthew Powell. Script by Caroline Kidd Story Editing by Levi Hanusch. Audio Engineering and Editing by Matthew Powell. Narrated by Matthew Powell. Video Production by Levi Hanusch. Brought to you in partnership with Lawrence Livermore National Laboratory. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Picture a single road looping the entire planet. No lanes, no exits, no stoplights. There's one car on the road. Then a few more cars. Easy to track. Easy to predict. Then on ramps open. Everywhere all at once, more cars merge on.

Then more. None of these cars have brakes. They have no way off the loop and nobody directing traffic. Then, a single crash. But the wreckage doesn't sit on the shoulder waiting for a tow truck. It explodes into a thousand new pieces with nowhere to go because every one of these cars was moving at 17,000 miles an hour. This is space. With objects fast enough to cross the United States in under 10 minutes and

140 million pieces of debris in its orbit. More objects present than people living in the country of Japan. On February 10, 2009, somewhere in orbit above Siberia, a Russian military satellite collided with an American communications satellite. At the moment of collision, each satellite was traveling nearly 36,000 kilometers per hour. And it's not just crowded. It's contested. This is an active, messy, military, commercial, public, private war zone that all of our economy runs through.

And that is a hard problem to ensure the safety and reliability of. A place where nations are watching each other, where satellites can be tracked, targeted, jammed, or deceived. How do we understand what's happening there? And how do we learn to operate in a domain that's moving? too fast for the old way of watching it. At Lawrence Livermore National Laboratory, one answer is being developed inside a mission operation center. One built for a space age that's no longer empty.

Welcome to the Big Ideas Lab, your exploration inside Lawrence Livermore National Laboratory. Hear untold stories, meet boundary-pushing pioneers, and get unparalleled access inside the gates. From national security challenges to computing revolutions, discover the innovations that are shaping tomorrow today. For most of human history, space was something distant we looked up at. A dark canvas that felt distant enough to belong more to imagination than everyday life. Wow. Then we learned how to reach it.

You are hearing the actual signals transmitted by the Earth circling satellite. of the great scientific feats of the age. It gave added incentive to the development of Earth orbiting satellites for photo reconnaissance. No single space project in this period will be more impressive to mankind or more important for the long range expiration of space. We have a liftoff, 32 minutes past the hour. We sent up rockets, satellites, telescopes, instruments, machines built to see farther and navigate better. Now, space is woven into daily life in ways most of us hardly notice.

I used a Maps app to get directions to come here, and none of that would work without space. Luke Peterson is the associate program leader for data science in Livermore's space program. When I order a package to get delivered, the logistics, routing, all of that, that all goes through space. Our financial system depends upon space. What happens when space stops being a backdrop and becomes a pressure point? When the same domain carrying our maps, money, and communications starts changing faster than the people watching it can respond. So this is very, very tough how you deal with it. It's a hard national security problem, but it's one where we would like to maintain freedom of movement, freedom of communications.

We would like our military to be able to continue to use space as it sees fit. Philip Rittmuller is a lead engineer at Lawrence Livermore and the principal investigator who helped build a system designed to navigate. this problem. There's an equivalence to Moore's Law that states every 18 months the processing powers of chips will double. Well, there's an equivalent for space where every two years, the number of objects in space is doubling. A secure, stable, and accessible space domain is crucial as challenges to the United States and our ally space capabilities continue to increase. It's happening really fast. It's a supply and demand thing because the number of space objects doubling every two years means there's a lot more people who need

operations centers, but operations centers usually are banned by people, so they can't scale as fast as the rest of the industry. Philip experienced the problem firsthand. It's what his team was faced with when integrating systems operations for the Pandora spacecraft mission. As we were developing that, we were trying to find mission operations centers, and the cost of it was so high, we're thinking this is something we can do ourselves. For our missions, which are small and experimental, we don't want to have some big fancy operation center. We can just have an operations here on the lab with just the capabilities that we need, and it'll be much less expensive. That idea became Starmock. The satellite, telescope, aerial drone, and remote sensing

mission operations center, or StarMock for short. StarMock gives Lawrence Livermore the operational backbone for a new generation of agile, lower-cost, small satellite missions, the place where experimental spacecraft and payload ideas begin moving toward real mission activity. But Star Mock wasn't only built to reduce operational costs. One of the big problems we see in space is there's this desire to move quickly and innovate, but you have this risk of actual operations going on. You're actually flying satellites. So StarMuck allows us to test new technologies, new ways of doing operations and demonstrating those capabilities to folks that they then are more willing to take on and adopt. So it's bridging that

valley of death between how you could change space operations to actually getting it out into the field. And so by having that test bed of StarMock, it lets us test out some of those technologies, break them in a safe space. And it is an extension of the laboratory ethos, but into outer space. We like to push the boundary on technology and science and how things could be re-envisioned. And so StarMoc is like our little hub of what the future of space operations could be. A safe exploratory place to move fast, break things, and learn how to improve in the process. And last summer, they did. In July 2025, a satellite company saw something they didn't expect.

One of their satellites was being flooded with commands. Hundreds of them in a single day. Far more than any had ever been designed to receive. But the signal wasn't coming from an adversary. The source? Three interns at Lawrence Livermore. We had been kicking around the idea of introducing a new way of automating satellite tasking. Our current satellites, if they want to take a picture of something, they issue the commands,

they send it to somebody else, and they said, okay, yes, this is the command we're going to issue the satellite. But it's very long manual process. And so what that means is that the time between, hey, have an idea, I want to take a picture of something, and when you actually get that picture, can be very, very long. Even in perfect conditions, the request itself would take at least 15 minutes. Then, between communication procedures and orbital positioning, the whole process from human to satellite to result could take hours. If we look at the space environment, particularly in the national security arena, waiting a few hours is just too long. If there's a threat coming, you want to know right away.

Star Mock asked a few students to spend a summer trying to solve the problem. We gave them the task of interfacing with an existing ground-based tasking. telescope network and extending that to the space layer. So the space force has an existing automation network for their ground-based telescopes, doing the problem of space domain awareness. And we said, well, could we just extend that to space? But pointing a satellite in orbit is not the same as pointing ground-based telescopes. The students had to write the translation between the two systems. And after three months of concentrated effort, the students changed. changed the time of an automatically tasked picture from a satellite of the International Space Station

from 15 minutes down to 10 seconds. It was so fast that we actually ended up breaking the satellite on orbit because in one day, I remember it, they sent over 700 commands to the satellite, and the satellite operators weren't prepared for that. We heard when we showed this to people at a big conference in the fall that the owners of the Space Force ground-based television, scope system, said, oh, wow, you did that so fast. We have private companies that have been trying to do this for years. Star mock gave them something most space experiments never get. A place to

try the idea on real systems, find the breaking point, and learn from it with speed. A difficult industry obstacle overcome in a single summer. And in today's space age, that speed matters. There are adversaries who realize our dominance in space. And they want to be a lot of space. And they They want to overtake it. They want to establish their own rules of the road. They want to establish their own norms. They're being increasingly aggressive out there. And so the space mission is becoming harder and harder and harder.

Orbit is active. And sometimes it's intentionally disrupted. The Space Force and other Pentagon agencies will need to respond to the growing threat of anti-satellite capabilities. A new report says Russia and China are making advances in counter space weapons faster than the U.S. is improving its defenses. Our space program really got started after 2007 anti-satellite tests by China, where they blew up a satellite in orbit, and it created a whole bunch of debris.

According to U.S. government officials, after three misses, China succeeded in shooting down one of its own aging weather satellites, with a medium-range ballistic missile fired from the ground. So our things in space are not necessarily immune. Our adversaries could attack it. There are countries who are pursuing very aggressive, very impressive counter-space capabilities. But an attack on a satellite doesn't have to mean an explosion. An adversary might want to interfere with a mission. It could be jamming your spacecraft or maybe they'd want to take a picture of your spacecraft, and you don't want them to know what's on your spacecraft.

You could teach an agent how to avoid that situation. To make that kind of autonomy possible, software has to be trained and tested far before it's launched into orbit. That's the kind of work StarMock, is trying to make possible. As you're developing your satellite and trying to decide if it works, we have the electrostatic discharge or ESD compatible room where we can put in our flight processors and our sensors and make sure everything talks and works together

and simulate the environment it will be in using HPC, high-performance computing. So we can test things out to make sure our hardware works, test the sequence of commands we might want to do, run through that, And then when we're in actual operations, things will go much more smoothly because we tested it all. This is what the team calls hardware in the loop testing. In that room, they can practice the whole mission before launching. So you could be running your flight processor with all the software that's going to run on the spacecraft, and it decides when to do a burn to change its orbit.

You can simulate that whole process. I want to do a turn and then a burn in this direction. That goes out to the HPC, which is running a simulation, and the simulation would say, okay, this is how your orbit would change. This is where you are now. And if you're looking in the direction you told us to look, you're going to see this.

Simulate the environment isn't just a metaphor. StarMoc sits behind the lab's firewall, connected to the computing power needed to recreate the world the satellite will face once it's in orbit. once it's in orbit. Inside that simulated world, engineers can test the satellite's brain before the satellite ever leaves the ground.

They can give it problems, watch how it responds, and refine the way it makes decisions. One of the tools that makes this possible is called SAPI, or Space Situational Awareness in Python, which was developed by LDRD or lab-directed research and development. We wrote this code that can run efficiently on high-performance computing, And what you can do with that is create scenarios where you have satellites flying around the moon,

that can fly around Earth, Jupiter, the drag from the solar wind. All this is like physics is in there. And so you get high accuracy modeling. And if you wanted to teach an AI to fly satellites, you can then give the AI that tool and say, okay, you can try to fly a satellite through our simulator, play your little video game and figure out if it crashes. And if it crashes, then change it and figure it out. You lose. That kind of simulator only matters if people can actually use it.

In space research, many of the tools that model orbits and satellite behavior are expensive proprietary systems. Sappy takes a different approach. Lawrence Livermore made it open source, giving students, researchers, and smaller teams access to high-fidelity space modeling software without needing a major commercial license. We open-sourced it, so it's free for the community. which is a big deal in this area because a lot of the other tools are proprietary and tens of thousands of dollars to run their code. That is actually what we did using our high-performance computers.

You give an AI this simulation tool and you let it play millions and millions of games and it eventually figures out how to fly satellites. Training that AI takes enormous computing power. Running it once trained takes almost none. We would like to develop games using AI and HPC that the Guardians can play to learn how to control a spacecraft and to win engagements. And we could use the StarMock for that. You could have two teams, one team of guardians in one room of the StarMock, another in another room. And they can play against each other and see what techniques work and what doesn't against the simulation that's being run on the HPC.

an entire training happening inside a simulated world. A satellite agent learning how to operate in a place it's never been. Then the trained agent moves onto a flight processor small enough to ride a satellite. That's how a spacecraft might act faster. But if satellites can start acting faster on their own, how does anyone keep track of the autonomous sky? Humans can't keep track of this anymore. It used to be that space domain awareness was a... human endeavor. There were 10 or 20 things you had to keep track of. Some operators in the US military

could once a day check in, be like, yep, that thing's still there. Great. You can't do that with tens of thousands or hundreds of thousands of objects. And so you have to automate. There's this big desire for automation, but all of the systems are legacy invented in the Cold War and many of them are still running horribly old computer languages and on very, very old systems. Things in space are described by what's called a T-L-E or two-line element. Why is it two lines? Because it's what the punch cards used. The Star Mock allows us to be like, okay, we need to move past punch cards.

But like how? Because we have all this other stuff we have to keep track of, and our operations are getting harder and harder and harder. There is no space to innovate in this area. If you have a Star Mock, what you can do is you can prototype these things. And be like, hey, actually, here's a way you could automate satellite tasking. We worked it all out. Here's a data schema.

it's not based on punch cards anymore. The sky now produces more information than any analyst can hold at once. Space Force alerts, satellite maneuvers, close approaches, breakups, each one a clue, but rarely the whole story. But there are threads between these pieces, and a Lawrence Livermore student prototype named Supernova was built to connect the dots. Imagine you have a chatbot in front of you, and you can ask it questions about

satellites. Hey, has satellite A, this U.S. satellite, does it have any relation to satellite B, this Japanese satellite? Ask the question and something assembles on screen in real time. In front of you on the chatbot on your screen is created in real time. It looks like a spider web. Or maybe it looks like that guy from the meme of like all of the red threads connected behind him. I've stumbled onto a major company conspiracy, Mac. How about that for stress? On your screen in real time that a conspiracy theory board gets created connecting the U.S. satellite to the Japanese satellite. And the chatbot comes back and says, yes, actually, there is a relation between these two. And there's this other satellite that has made close approaches to each of these independently.

Months apart, there's this other satellite from China that has actually gone near both of these. That's the kind of information that an analyst would really like to know. it would take them months and a fair amount of luck to find the connection between U.S. Satellite A and Japanese Satellite B. One alert may only matter because of something that happened months earlier. The future of space operations depends on seeing those connections while there is still time to act. StarMoc's new technology is as much about understanding relationships as it is tracking objects. The future of space exploration isn't being built around one-two-term.

one student project or one mission. It's an entire infrastructure, a place where new ways of operating can be tested before anyone knows exactly what they will become. I think of it like this. When DARPA developed the internet, they didn't know that it was going to enable Google and smartphones and Amazon. They developed this infrastructure and put it out there and let super smart people figure out what to do with it.

That's what the Star Mock is. We've developed infrastructure and I'm really excited to see where it goes. And it's already reaching beyond the lab. The Space Force is standing up an AI accelerator centered at Stanford and Livermore is signing on as its mission transition partner. The place where a promising idea from academia gets pushed from ideation into real mission. I want to put it out there and let this team of super smart people we have here at Lawrence Livermore go tackle new problems and figure out what they want to.

to do with it because I am sure that I'm going to be surprised with some ideas that come out of this. Above our heads, the traffic keeps merging on. Faster satellites, smarter adversaries, a sky doubling in objects every two years. Whoever learns to fly it, to teach the machines to read the whole crowded lane at once, to move in seconds instead of hours, will help set the bar for a safer, more secure sky. Thank you for for tuning in to Big Ideas Lab. If you loved what you heard, please let us know by leaving a rating and review.

And if you haven't already, don't forget to hit the follow or subscribe button in your podcast app to keep up with our latest episode. Thanks for listening.