Standing between satellites and space junk: Robotics, and a helicopter pilot
Marco Peterson is a fourth-year graduate student in the Kevin T. Crofton Department of Aerospace and Ocean Engineering. His work combines machine learning and robotics to design systems that can perform complex tasks, including orbital satellite repair and space station construction. Full-time research isn’t Peterson’s only job: He’s also a reservist in the U.S. Army, where he’s a CH-47 helicopter pilot with more than 300 combat flight hours and company commander of Delta Company 5-159th; he paused his graduate work in 2019 for a one-year deployment to Iraq. Peterson’s research is supported in part by the Institute for Critical Technology and Applied Science through its Doctoral Scholars Program. His doctoral advisor is Jonathan Black, a professor of aerospace and ocean engineering and the director of the Aerospace and Ocean Systems Laboratory at the Ted and Karyn Hume Center for National Security and Technology; Black is also the co-director of Space@VT.
Interview has been condensed and edited for clarity.
Your first graduate degree is in computer science. How did you wind up in aerospace engineering?
After my master's degree, I went to the Army's flight school to become a helicopter pilot, and near the end I was like, well, now what? I'm a reservist, I'm not on active duty. So I just started submitting applications to graduate schools to see what would happen. Then one day, after PT in the morning, I get a call from a 540 number and it was Dr. Black saying “Would you like to get a Ph.D. at Virginia Tech?” I said, “Let me think about it. What's your computer science program like?” And he said, “Oh no, no, no, it won’t be computer science — it'll be aerospace engineering.” I accepted the offer and the learning curve for my first semester here was just about straight up.
Why do we need robotics in space?
When your satellite dies in orbit, it becomes a piece of space junk. If there's a team of robots out there that are sitting in specific orbits, you can redirect a robot to see what’s wrong with the satellite or execute a repair mission.
If I had to put it in a sentence, I would say, think of R2D2 and that's what we're trying to do. The crux of the research is using computer vision and artificial intelligence to autonomously identify objects in space. There are a lot of variances: Orientation, because there is no down, and lighting, because you can be on the dark side of a station or a dark side of a celestial body.
What do you hope the outcome of your work will be?
We've been sending humans outside of airlocks to perform repair missions for a long time and are pretty good at it. But if I’m the guy sitting down in Houston, going through all the scenarios of what could go wrong, I'd much rather lose a robot or a piece of hardware than a human. So what we're trying to do is replace that human in the loop with a robot in the loop. Once this problem is solved the use cases are endless. As access to space becomes cheaper and cheaper, you'll start seeing larger and larger structures being assembled, like a giant spaceship being assembled in a dry dock.
What does a typical day in your lab look like?
Right now I'm trying to get a single-board computer to talk to my drone correctly. We've done this before with something called a Raspberry Pi. However, in order to do neural networking and image processing, you need a dedicated graphics card, which the Raspberry Pi does not have. So we've upgraded to another single-board computer. But that computer has a communication protocol that is not playing nice with my current architecture. So right now, what my lab looks like is staring at a computer screen, figuring out why the code doesn't work.
Graduate school is a full-time job. How do you balance that with your military career?
My life is pretty much a constant state of chaos. For most reservists, the commitment is one weekend a month. But as an aviator, you have to maintain 7.5 flight hours a month, so I'm usually at Fort Eustis four or five days a month getting flight hours in. I'm also a company commander for the largest aviation maintenance company in the U.S. Army. On both sides I have very competent teams. At Virginia Tech I’ve built teams for both the machine learning piece and the robotics piece, including another doctoral student, a master’s student, and several employees and undergrads from the Hume Center. I'm in the trenches with them. But whenever a task needs to be done, it's pretty well organized as to who's responsible for what.
What was it about space research that attracted you?
This is one small step toward becoming an interplanetary species. That's the last direction we have left. We've explored everything else except for maybe the Marianas Trench.
I've always been interested in space. I've applied to the astronaut program twice. Part of my master's degree was an externship at NASA Langley, and of course everybody there applies to the program. The running joke is that if you get denied — which all of us do — you need one of two things: More flight hours or more degrees. I'm working on both.
What’s next for you?
After I graduate, I'm going back to the military side for about a year, attending a maintenance test pilot program. I always tell people that flying is the best job in the Army. My dad was an F-15 mechanic, so I guess you could say I'm the second generation of aviation. Every branch of the military is short on pilots, but they're especially short in the maintenance test pilot area. My command has asked for volunteers and I've always wanted to do it. Then I'll start looking for jobs.
What advice would you give an incoming graduate student?
Please invest in a solid computer. Your first year you're gonna have to run all these programs in MATLAB. Bring in a solid computer. I'm an avid gamer, so I've always had the hardware to do what I needed to do. But I spent most of the day yesterday getting computer hardware working for some other teams. So I'm always pushing for more computers.
The ICTAS Doctoral Scholars program helps attract exceptional students to Virginia Tech and increase the educational and experiential diversity of research by offering a competitive graduate fellowship and professional development opportunities to students pursuing transdisciplinary research. The program is led by ICTAS and supported by significant contributions from the Graduate School and each scholar’s college and department. Faculty members interested in nominating an incoming graduate student for the program should contact their college’s associate dean or graduate director; more information is available on the ICTAS website.