Embry-Riddle Student Team Earns First Place in Capture-the-Satellite Competition

Embry-Riddle XDLab Team wins third annual capture-the-satellite competition at AIAA SciTech Forum 2026. (Photo: Courtesy of Aerospace America/David Becker))
Embry-Riddle's XDLab team won the third annual capture-the-satellite competition at AIAA SciTech Forum 2026. (Photo: Courtesy of Aerospace America/David Becker)

A team of Embry-Riddle Aeronautical University students has taken the top spot in a national, months-long capture-the-satellite competition, a game that teaches how to control a spacecraft to avoid obstacles and safely rendezvous with another spacecraft that needs service.

The group from Embry-Riddle’s XDLab beat out 51 teams and two other finalists to earn first place at the competition held Jan. 15 at this year’s AIAA SciTech Forum in Orlando. The competition features a problem in which students must determine how to maneuver, approach and capture a malfunctioning satellite, even with obstacles in the way.

Team co-captain Tyler Barr, an undergraduate Aerospace Engineering major, said the exercise was extremely helpful, as he was learning in class many of the theories needed for the game. 

“There are so many different sides to it. There’s the software side. There are the orbital mechanics. I was learning in class, at the same time I was learning this. So I was able to apply the stuff from class to this competition. And I thought that was really cool,” Barr said.

The Embry-Riddle team was mentored by Dr. Di Wu, assistant professor of Aerospace Engineering and director of XDLab, which looks to integrate space engineering and artificial intelligence.  

“There is increasing interest in building an intelligent, sustainable and vivid space industry,” Wu said. “And these students had the opportunity to build their skills, as they inevitably are going to be involved in integrating AI and space.”

This year, students had four months to work on a “lady-bandit-bodyguard problem.” In this situation, the teams had to use software and algorithms to navigate and control a satellite, called the bandit, that had to contend with another satellite (the bodyguard of the target, chasing the team’s satellite) — on their way to rescuing the target, or lady (the satellite), explained Ross Allen, technical staff member in the Tactical Autonomy Group at MIT Lincoln Laboratory, who runs the program.

 “We’ve tried to generate a really challenging, difficult problem here,” Allen noted. “From the participants’ perspective, they’re in control of one satellite and they are interacting with two other satellites. They’re trying to pursue this malfunctioned satellite while evading a third satellite, all operating in close proximity.”

 Surprisingly, the Embry-Riddle team did not rely solely on an AI approach but used a novel hybrid method combining classical orbital mechanics and AI.

“We thought just using AI and (machine learning) might be risky, because it’s hard to explain what the AI is telling you to do,” said Luca Sportelli, an Aerospace Engineering Ph.D. candidate and team co-captain.

There has been a longstanding battle between the classical and AI approaches, Wu said.

“Yes, we’ve seen teams diving deep into one way or another,” he said. “However, it should be about distilling a combination of the two approaches. This is not only applicable to space systems, which XDLab focuses on, but would have further implications for the incoming trend of integrating AI into any system.”

Allen said the exercise helps spur research and development of algorithms and AI agents for autonomous control of spacecraft. The platform is based on the Kerbal Space Program, which provides a full-featured physics engine for low Earth orbital mechanics. The problems the students are solving can be applied in the real world, especially as the number of satellites in orbit increases.

“I see it as very much a kind of real-world experience developing autonomous systems. In particular, it’s giving them experience working on a very nascent problem of space autonomy,” Allen continued. “The underlying algorithms they’re developing that would drive that control system, the underlying theory, would be highly applicable to the real world.”

Teams started working on the problem in August 2025, with the semifinal round held in December. Students had three days to come up with a program for the finals.

“They’ve had all this time to figure out how to outsmart me, how to outsmart my software,” Allen said. “But then I released something new that they haven’t seen, this third satellite. I was hoping that they’ve created something that is more robust, that’s not really brittle and fails as soon as it sees something new.” 

Indeed, that’s exactly what the Embry-Riddle team did, evading the guard satellite and capturing the lady target.

Embry-Riddle contributed to and edited the content of this article for university purposes. The original version [https://aerospaceamerica.aiaa.org/institute/gotcha-students-capture-malfunctioning-satellite-in-spacecraft-control-exercise/] was published by Aerospace America and reported and written by Lawrence Bernard.