Alumni Spotlight: Josh Vander Hook
Working as a systems researcher at NASA’s Jet Propulsion Laboratory (JPL), Josh Vander Hook uses his robotics expertise to look at what technological possibilities are on the horizon for future missions that will help NASA put, first, robots and, then, humans on Mars.
As part of a large team working on the exciting Mars 2020 mission, Vander Hook is helping to build on discoveries made by other notable missions, such as Viking 1 & 2, Pathfinder, and the paired rover mission of Spirit and Opportunity. He has brought his Ph.D. work on cooperative robotics to JPL to address the distribution of computing and communication tasks over a proposed ecosystem of rovers and satellites that will one day further explore the red planet.
But before Vander Hook found himself looking up toward our celestial neighbors, he was waist-deep in Minnesota lakes with advisor Volkan Isler researching invasive fish species for the Robotics Sensor Networks lab (RSN).
How did you go from doing graduate work here with CS&E to working on a high-profile mission for NASA’s Jet Propulsion Laboratory?
While with the Department of Computer Science and Engineering, I basically won the robotics lottery. I got the opportunity to work on a real-world application using multiple cooperating robots. The project involved building boats that could be dropped in a lake and track down radio-tagged invasive fish. It was a great intersection of science (the fish ecology), engineering (the robots), research (the important guidance, navigation, control, decision-making algorithms), and a societal need (the infestation of carp).
My advisor, Professor Volkan Isler, and members of my committee helped me make initial contact with the JPL, and when I visited, it just felt like a good fit. I think it was because NASA (and in particular JPL) pursue the same balance of science, engineering, research, and societal need as Volkan’s lab. In fact, there was a major project at JPL with a need for building autonomy into boats and submarines for the Navy, Marines, and other applications that was similar to my Ph.D. work. I owe it to my experience at RSN and the University of Minnesota for preparing me to work on that project.
How are you applying your Ph.D. work into the exciting Mars 2020 mission?
I worked briefly on the Mars 2020 path planning software. The Mars 2020 rover must travel much further, and faster, through rougher terrain than previous rovers. My current project involves researching the impact of new processing capabilities into future Mars missions (starting with 2020), and what new capabilities and mission objectives might be possible given advances in technology, in particular new computing architectures. I think the answer is to stop considering each new Mars-bound asset as a single robot, and start treating the whole ecosystem as a “system.” This teases out a lot of interesting questions in the interplay and redistribution of computing tasks and communication requirements. For example, is it cheaper to use three systems that are always linked or give two systems mobility so they can rendezvous when they need to share data?
The mission is a good fit for me because my Ph.D. work involved a ton of classic algorithm analysis, systems building, and asking the question: What if we automated it? We in RSN spent a lot of time thinking about these problems, and my Ph.D. was one of many in the lab that pursued these kinds of questions.
My initial interest in CS was theory. I loved self-optimizing data structures, algorithms with provably good performance and runtimes, and geometric algorithms in particular. But there is something addictive about building things that I couldn’t shake, and when I first saw a robot running software I had written, which combined all of my previous interests nicely, there was no turning back. Along the way, I stayed with robotics for various reasons, mostly because of the promise and potential for assisting humans with tough tasks. We see it when robots helped us explore the Fukushima reactor, other planets, or the oceans. The only interstellar spacecraft we know of is our robot (the Voyager probe). Robots will very likely be the first to find extraterrestrial life!
But another big point is that robotics is such a broad field. You can do any sort of classical engineering under the “robotics” banner if you are also willing to roll up your sleeves and test it on a real-world system, with a real-world application in mind, and in real-world conditions. Often, this involves snow, sand, sun, cold, heat, and other interesting challenges that you don’t appreciate when sitting at a desk. That’s really the big reason I love robotics: the elbow grease.
What inspired you to pursue computer science as a career?
How could one not be inspired by computer science? Computers are (in my opinion) the most recent of our society-changing inventions. Before studying CS, they were a mystery. A device with a computer in it was a magical item with an unknown purpose, capable of all kinds of things. Want a greeting card that can sing when you open it? No problem. You can get into programming with almost nothing but patience and a second-hand desktop. I stuck with it as a career because if I didn’t, I’d have trouble finding time to do other work while I was tinkering with all these computers.
Do you have any advice for current Ph.D. students who are looking to follow a similar career trajectory?
I can only speak from my experience, but the field of CS, and especially robotics, is so huge. It’s impossible to know where to dive in, or what to study. The answer is anywhere and anything. It’s OK to trust your instincts and do what fascinates you. You don’t have to make the right decision from the get-go, as long as you keep an open mind to where the path leads. Along the way, you must seek inspiration and advice from others. The myth of the solo inventor is hard to shake but human knowledge has eclipsed what a single human can grasp (at least for me!). It’s important to talk to a lot of other folks who can distill their knowledge for you, help you find good directions, and, most importantly, find a place to make a real impact.
What are some unique challenges you face working on a project intended to lay at least some of the groundwork for how a future human explorer could live off Martian resources?
Where to begin? The first is the pressure. I share the load with literally thousands of more-capable engineers at JPL, but still, in the back of my mind I am always thinking “Is this the right direction?” or “Am I spending your tax dollars correctly?”. There’s no sense pushing my pet project if it will not be used or won’t inform future missions. The U.S. places incredible trust in its researchers (in the form of funding), and that’s important to keep in mind. The main challenge is distilling myriad ideas into actionable directions quickly and efficiently. We take our orders effectively from the citizens of the U.S., through congress, as informed by scientists, but it boils down to questions like, “How do we get cameras and salinity sensors through kilometers of ice, on a moon 1.2 million kilometers from us, so we can sample the ocean below for life?”. That’s a tough nut to crack. (For the record, a colleague of mine thinks autonomous robots climbing down cryo-volcanoes is the answer. How cool is that?)
What is it about computer science that keeps you excited and motivated for the future?
I think CS has such an impact on science and engineering because of the ubiquity of computing. We won’t put the genie back in the bottle now, so the reach of CS will only increase with the arrival of more automation, AI-driven analysis, next-generation IoT, and all the software magic that keeps it running. I’m excited to see how we change our interaction with computing systems. Will we find them to be trusted advisors and aggregators? Will we see a future of smart-devices in which the software is largely hidden? Will we use cryptography to rebuild privacy and expand trust in institutions? What policy decisions could be affect by the inclusion of AI-driven analysis and is that what we want? What new computing architecture will cause a revolution in CS? It’s a fascinating time.