By Deb Derrick
Maybe he was destined to an engineer. As a young boy growing up in central Nebraska, Shane Farritor put together spaceships and bulldozers with Legos and Erector sets. He ran loose in his fathers hardware store like a kid in a candy shop. Destiny or not, Farritors interests steered him toward mechanical engineering at University of NebraskaLincoln, where he graduated in 1992. He went on to study robotics in graduate school at the Massachusetts Institute of Technology, receiving his doctorate in 1998.
Now Farritor has come back home to Nebraska. And hes still tinkering with the same kinds of ideas he had as a childwith the same level of enthusiasm. Farritor talks about his work in this interview with Contacts.
|Q: Robots are cool. How long have you been hooked on machines?
A: I was building spaceships with Lock Blocks when I was three years old. My dad owned a hardware store in Ravenna. I made all sorts of things with Legos, Erector Sets and stuff around the store.
Q: You went to UNL and got interested in space and planetary exploration first.
A: As an undergraduate student, I did a co-op at NASA Goddard Space Flight Center in their mechanical testing branch. I also had a summer internship in the Space Life Sciences Training Program at the Kennedy Space Center.
Q: When did you start doing work in robotics?
A: As a masters student at MIT. I was a researcher at Draper Laboratory from 1992 to 1995. We designed and built a few unmanned robotic vehicles like the Sojourner rover that NASA used on the Mars Pathfinder mission. Our rovers were called MITY rovers. There were three of them, but I only worked on numbers 2 and 3.
Q: What happened to these rovers?
A: Theyre being used for landmine detection now. One of the most innovative features of the rover is a sun sensor. On Mars, you can use the sun for navigation. A compass doesnt work because the magnetic field is too weak. So we took a video camera, pointed it toward the sky using a fish-eye lens from a door. Then, you assume the sun stays in the same place. As the rover turns, you watch how the sun moves to see which direction you are traveling. The last I heard, NASA thought it was a good idea. Its been on the last three research prototypes for the Mars 2003 rover.
Q: So you were having so much fun that you didnt want to quit?
A: It was a lot of fun. I did my doctoral work on modular field robotic systems at MITs Field and Space Robotics Laboratory. My adviser, Steven Dubowsky, is highly regarded in the field of robotics and has worked with NASA for a number of years. I studied reconfigurable robots for use on future Mars missions at the NASA/Cal Tech Jet Propulsion Laboratory one summer. Also, between my masters and Ph.D. work, I spent about nine months in the machine shop building a mobile robot called Companion.
Q: What does Companion do?
A: Companion was originally designed for the Department of Defense to help soldiers in the field with tasks like carrying ammunition. It was built on an electric wheelchair platform and its fully autonomous, meaning it can go places on its own. It has proximity and contact sensors. It also has a ring of acoustic sensors to detect obstacleslike a bat. It has two cameras. It also has a laser that can detect the distance to various obstacles. Its a very capable robot.
Q: Now you want to use it for highway maintenance tasks.
A: Weve been talking to the Nebraska Department of Roads about this. There are a lot of jobsgrinding, sealing cracks, bridge inspection, patching potholesthat are dirty and dangerous. These are good jobs for robots in terms of safety, and they could also save money.
Q: Youre also working with Kawasaki.
A: Were trying to help improve the way their robotsthe ones they manufacture at their Lincoln plantare programmed, to make the process safer and quicker. The current programming method requires the operator to use a remote control, called a teach pendent, to move the robot to every desired location. Were building a new sensor system so you just need to move a small device to the desired location, not the entire robot. This is similar to how you move a mouse to operate a computer.
Q: What about any projects with planetary exploration robots?
A: Ive had a couple of grants from NASA to study a modular robot design that could be used on future missions. NASA will send two large scientific rovers to Mars in 2003, and humans will set foot on the planet someday. Robots will be needed on these missions to do a wide variety of jobs. Theyll unpack, assemble and deploy equipment, collect samples and many other tasks. Each of these tasks poses a significant challenge for robots as theyre now designed.
Q: How does the design pose a problem?
A: Robots are typically designed to perform a few specific tasks. They might be capable of moving and deploying a scientific instrument, for example. But they cant dramatically changed to
|do other things. Once their job is done, you cant easily use the hardware for other purposes.
Q: How could they be designed differently?
A: Weve come up with the idea of a modular infrastructure of robots, rather than individual robots. The system consists of modules that are assembled to produce a robot for a specific task. The modules include actuated joints, links, end-effectors, sensors and mobility units. You can put them together into dramatically different configurations for various tasks.
Q: Isnt NASA looking at this already?
A: Not many people are looking at human Mars exploration at this level of detail. NASA is more focused on short-term goalslike the space station. Weve come up with designs for nine robots that could be put together with only 26 modules. Then we used computer simulation to demonstrate their ability to perform mission-relevant tasks. We simulated a robot digging for a science sample. We also simulated a robot deploying a scientific instrument and another one that lowered a front loader, moved forward and pushed a boulder. Based on our simulation, the concept is scientifically feasible. But theres a lot of work to be done.
Q: What technical challenges are you facing with this modular robot design concept?
A: There are a lot of technical and theoretical challenges associated with modular design. Im interested in modular design theory. Modular design is, in many ways, different than traditional design. In modular design you are given a box of parts and asked to put them together to do a specific job. Kind of like the Lego problem. There are a lot of interesting questions in this problem. If you remove one Lego from the set, does that greatly reduce the number of solutions you can produce? What if you added one super Lego, could that increase the number of solutions? I think its interesting.
Q: Where do you want to go with this project?
A: I hope all the way to Mars.
Q: Will we see robots like Rosie on The Jetsons or Data on Star Trek?
A: Its going to be a long time before we get to the Rosie stage. Soon well probably see robots help around the house with chores like vacuuming and mowing. People are already thinking of strange applicationsits unbelievable what crazy ideas are out there.
Q: Robots seem to appeal to kids of all ages. What excites you about robots?
A: Mechanical engineers are hands-on people. They make things happen. They do things. Robots do things, too. Going from theory to lab work is a big step. Its a lot harder to build a robot than to talk about it. You can see what a robot does and understand whats going on. To me, thats more exciting than seeing something on a computer screen. Thats the kind of excitement I hope to share with our students.
Q: I understand MITs robotics competitions are big events.
A: They go nuts over this competition just like we do over Nebraska football. Youd have to see it to believe it 500 to 600 students crammed in a big roomstanding room onlyyelling and screaming as these robots go head to head. Its pretty neat. The MIT competition was the precursor to the BattleBots TV show.
Q: Now that bots are on prime time TV, how about doing something here in Nebraska?
A: Stay tuned.
For more information on Dr. Farritors work, go to <robots.unl.edu>.