Jimmy Ragan - Three Minute Thesis - 4/25/2023

if there's one thing I've learned since moving to Southern California is that it's important to keep an eye out for bad drivers I was taught to drive defensively that I can't assume others will be careful that my safety was my own responsibility why can't this be the same for our robots in the not so distant future swarms of spacecraft will work together to construct space stations

that's where my research comes in I'm teaching spacecraft to drive defensively and I'm doing so by giving them the tools they need to anticipate and better respond when things go wrong on the road when interacting with other drivers we make predictions and assumptions about how people behave spacecraft can do something like this too using knowledge of the spacecraft's hardware and what the Swarm is trying to

achieve we can predict how each spacecraft's neighbor should act when everything is working versus how they might act when things are broken for example let's imagine we're trying to join a lineup of spacecraft our neighbor should slow down and merge behind us but if their thrusters are broken they might be unable to do so we can start to anticipate these future dangers by thinking several steps

into the future but just like in a game of chess these future interactions get very complicated very quickly to handle this I like to think of the future as a tree of possible outcomes branching every time our spacecraft or their neighbors make a decision these trees can get very big but thankfully we don't need to consider every possible future just the ones where we might be

in danger or where we might learn something about what is wrong and how we should react by focusing on the most important possibilities we can plan far enough into the future to act now in a way that gives us the best chance of staying safe such as realizing our neighbor is on a collision course and that we should slow down and make some space even if

we technically had the right-of-way it might work so far I've taken this idea and developed the mathematical model or framework to translate this into a form that computers understand and in doing so I've made the process fast enough and accurate enough to run on real systems for the rest of my thesis I'll be working with JPL to apply this approach to upcoming missions just like learning

to drive gives teens Independence and learning to drive safely gives their parents Peace of Mind enabling spacecraft to act independently will allow for more ambitious and complex missions that are possible today and giving the spacecraft the ability to anticipate and react to unforeseen failures will make doing so safe enough to be a reality thank you [Music] [Applause]