It turns out that controlling a $100 helicopter using only your hands and a LeapMotion controller is pretty difficult.
This was the lesson I learned after about 20 minutes of trying to get a quadcopter (a helicopter with four rotors instead of one) to fly around a carefully fenced-in area of the GDC Atrium without crashing into the poles, or worse, into my classmates. The project was a demonstration of the work being done by Matt Broussard and Robert Lynch, both ‘16, in Dr. Peter Stone's lab. The lab, which mostly focuses on robotics, has a fairly large segment dedicated to quadcopters and their study.
Dr. Stone's lab is an extension of a program known as the Freshman Research Initiative, or FRI, which gives about 500 freshmen the opportunity to do research in one of several labs across the College of Natural Sciences. For example, if your area of interest is astronomy, you can take Dr. Shapiro’s Cosmic Dawn stream, which introduces you to astronomy research. In general, the larger a field is, the more streams it has. Biology, for example, has a whopping 13 streams, whereas astronomy only has one. As the fields develop, more and more streams are added; for example, CS started with one stream a few years ago but has since added on another three streams in different areas.
Dr. Stone's lab is arguably one of the most popular streams in the program, since—I mean—you’re programming robots all the time. One of the lab’s most recent projects involves updating the CS electronic billboards, the touchscreen monitors in the hallways of the GDC, to include a (currently inactive) button to call a robot to your location to help you find, for example, the closest restroom or a particular office, as part of a program called Building-Wide Intelligence, under the supervision of Dr. Stone. Students are almost entirely responsible for the code that gets pushed out, and according to Broussard, “[the best part of my work] is that unlike many other things we do in CS, here the code we write actually has physical effects: a robot or quadcopter actually moves and acts!”
Of course, the most important part of any good research project is being able to actually do it (as opposed to, for example, being forced to do grunt work for grad students or postdocs). Both Dr. Matteo Leonetti, the lab's supervisor, and Dr. Stone were emphatic that they want students (especially freshmen) to get involved with their research, and according to both of them, the best way to get started is to take their CS 378 section in the fall through FRI. (Check out the CNS FRI page for more information on prerequisites, including Research Methods.) The course is designed for anyone, including people with zero coding experience; it’s designed to give anyone interested in robotics a chance to try their hand at it. In the spring, after the first few weeks of learning Python, students are usually given a lot of freedom to pursue their own projects. According to Broussard, “we've had a lot of freedom to just use [AR.Drones] as an explorative platform for developing techniques that might be useful on more practical platforms, as well as to create fun demos in the atrium that give us an opportunity to let people know about the cool stuff the BWI group is working on.”
Edit: I said in the article that this was Dr. Leonetti's lab, which was a mistake; the lab is Dr. Peter Stone's, but it's overseen by Dr. Leonetti. Also, students in the lab are responsible for programming robots, not building them. Sorry about that.