Automation is used in several ways. Could cleaning up space be the next?
Madhur Tiwari, aerospace, physics and space sciences assistant professor and director of The Autonomy Lab, will be working on the project with a startup company.
The research will explore spacecraft automation by using high-fidelity simulators. While the project will focus on space debris cleanup, Tiwari noted the technology could have other uses, such as automated repairs, docking or when multiple spacecrafts are functioning together.
With more spacecraft in use, there is a greater need for machines that can execute tasks without supervision. To meet that demand, researchers such as Tiwari are studying the location of these machines once they are in space and how their location affects their effectiveness at executing tasks.
“Let’s say I’m 10 miles away from you. I could be 10.1. I could be 10.5. I’m actually never exactly 10, right? This becomes a problem if you think about cars very close to each other,” Tiwari said. “This becomes a problem when spacecrafts are very close to each other. So we don’t want to be very uncertain.”
The contract will allow them to figure out how to assess, in real time, the uncertainties of navigation and operations.
“We want to quantify that uncertainty. We want to know, okay, I think we are in this bubble. We are in this range,” he said. “We want to be sure where we are in this range so we can plan other parts of the mission properly.”
This research also starts a new era with The Autonomy Lab. With a goal towards machine learning-based robotics and control, the lab also wants to make drones, quadcopters and spacecraft do things more independently. Previously based out of the university’s Center for Advanced Manufacturing and Innovative Design, the lab is now in the Olin Physical Sciences building in a space formerly occupied by the Orion Lab. Tiwari has a total of 29 students, including eight in the masters’ program and three doctoral candidates.
Tiwari grew up seeing NASA’s space accomplishments and being captivated by the imaginative worlds of Star Wars, Star Trek and other sci-fi entertainment. So the merging of machine learning and aerospace was a natural step for him. Couple that with an interest in the human function, and his research goals are about a natural evolution of engineering and discovery.
“I’ve always been intrigued with emulating how we as humans think,” Tiwari said. “We are such perfect machines, basically, in terms of how we want to do daily tasks, achieve our goals, and navigate our surroundings. We are so good at that. That’s where the machine learning part comes in. So, that’s the most exciting part for me, that striving for how close can we get to making machines like humans.”