Imagine Magazine - Johns Hopkins - May/June 2010 - (Page 38)

exploring career options Robotics Engineer Arin Morfopoulos Robotics Engineer, NASA’s Jet Propulsion Laboratory Arin Morfopoulos was a freshman majoring in computer science and electrical engineering at UCLA when he joined a pickup soccer game. An engineer from NASA’s Jet Propulsion Laboratory happened to be playing—and happened to need a summer intern. Morfopoulos took his programming skills to JPL in the summer of 2001—and he’s been there ever since. It might sound like a career started by luck, but as he explains here, what really opens doors in his department is the ability to make things. What kind of work did you do as an intern? We were working on an urban search-and-rescue robot, and I tested a wide range of things, including a laser range finder that the robot used to figure out how far away things were. The software that controlled the laser didn’t quite work right, so I rewrote part of it to fix a couple bugs. We were focusing on obstacle avoidance, so we would run the robot through an obstacle course, and we’d do test after test after test. Then we would tweak the programming and test again. But that was only for a summer. At the end of the summer, I’d done well enough in the lab that they wanted to keep me. So I continued as a student intern for the next four years in the same department, but on a lot of different projects. And then I was hired full-time when I graduated in 2004. Interview by Melissa Hartman software, it’s literally just sitting and thinking. No tires are moving. It’s just sitting there trying to figure out where the obstacles are in front of it and how to avoid them. So I’m working on something called a fieldprogrammable gate array, a single chip that will perform all the calculations the robot needs without using the processor at all. With this chip, the rover could figure out where the obstacles are quickly enough that it would never need to stop and think. What do you think is the coolest project you’ve worked on so far? Probably the Formation Control Testbed, which was built to develop technology for missions such as the Terrestrial Planet Finder effort to find Earth-like planets around other stars. For such missions, multiple telescopes will fly in precise formation to create a single image accurate enough to show something as dim and as small as a planet in another solar system. To prove that that level of precision flight in space is possible, we’re using robots that float on air bearings above a perfectly flat floor to simulate the frictionless conditions in space. These robots are 800 pounds and sixand-a-half feet tall, but because they float on a cushion of air, they can be pushed around the room with your finger. They are huge—and very slow. We call them the Blue Whales instead of the Blue Angels when they’re flying around together. What are you working on now? A few things. One is related to a problem the Mars rovers have with perception and navigation. This is one of the great unsolved problems in robotics research: How can a robot know what the environment is around it? Take a building for an example. How do you get the robot to recognize that it’s a building, that it can’t climb it or drive through it? Or how do you get it to know that a hill with a certain steepness is okay, but beyond that steepness is not? Or that steps are okay but a ledge is not? All those things that humans understand really easily about their environment are really hard to teach robots. When we give a command for the Mars Exploration Rover to drive, say, 100 meters in front of it, it first has to take pictures and run software to avoid rocks or other obstacles. It’s running on a processor that’s about 20 years old, so that command can take anywhere from 60 seconds to five minutes to process. While the rover is running that What is a typical day like for you? My work tends to be separated into two categories. There are times when I’m doing research, just sitting by myself, writing code. And then there are times when I’m working on electronics, and this is when I work with other people, like software and mechanical engineers, to make a whole robot work. 38 imagine May/Jun 2010

Table of Contents for the Digital Edition of Imagine Magazine - Johns Hopkins - May/June 2010

Imagine Magazine - Johns Hopkins - May/June 2010
Table of Contents
Big Questions
In My Own Words
If They Build It
Design That Matters
Wired to Win
A Student at SPAWAR
Even When the Ground Shakes
Biomimicry
An Engineer in Training
Engineering My Future
Home Away From Home
Selected Opportunities & Resources
Middle Ground
Off the Shelf
Word Wise
Exploring Career Options
One Step Ahead
Planning Ahead for College
Students Review
Creative Minds Imagine
Sudoku
Knossos Game

Imagine Magazine - Johns Hopkins - May/June 2010