Flying Autonomous Machines
Associate Professor Ella Atkins speaks about flying autonomous machines at the University of Michigan. Atkins works on some unusual autonomous unmanned aircraft, such as a recent collaborative project on an unmanned seaplane called Flying Fish.
Transcript
00:00:07 My name's Ella Atkins, I'm an associate professor in aerospace engineering department some of the things that i've worked on includes building autonomous unusual unmanned aircraft such as a... collaborative project with some of the other professors to build an unmanned sea plane called flying fish it ended up being the first fully autonomous sea plane ever built where it not only can fly uh... planned trajectories but it also could decide when it took off and where it landed my particular area of focus is on autonomous systems looking at how i can either make the system behave in a more intelligent way
00:00:43 or have it handle anomalies which is a big issue in air space right now so let's say your engine fails, so if you look at something. One of the things I studied U.S. Airways flight was hit by geese in the new york area. You can calculate that, that plane had the automation, had the capabilities which is very reasonable given todays technology to select a landing site and immediately return. It turns out the best thing would be for it to have returned to Laguardia. It could have landed. The pilot did not have the ability to make those kinds of mathematical computations even though they were relatively simple in the short timeframe. Computers can do math really fast. So if they understand the nature of the problem, they can solve the problem in
00:01:23 less than a second. Whereas the pilots really have to spend a while just trying to understand what the problem is. At a university what we're really looking at is ways to improve the technology through research. So here we're doing a variety of things. This is a platform that has been developed to allow testing of flapping wing vehicles in a wind tunnel vacuum and still air environments. What we're looking at is to better understand the airodynamic forces and our ability to sense them on a small scale platform. The sensors on this include a load cell which is capable of measuring all the
00:02:01 forces and torques acting on the system pressure ports that are embedded into the wings themselves that allow us to measure in real time the pressure distribution over the top and bottom surfaces as this flaps at different frequencies under different atmospheric conditions ultimately what this will allow us to do is to validate the be computational models of flapping wing platforms which are not well understood. This allows us to actually develop realtime estimates of lift and drag that are not just kind of based on the models that you have before you fly.
00:02:32 Right now I think the public sees them as tools of the department of defense. Anytime you can send a camera or another sensor into a hostile environment in place of a human, that's a win situation because the human isn't put at risk even if we lose the unmanned aircraft. Law enforcement has a number of applications having the officers aware of what's going on in a hostile urban environment, in a disaster situation. So for example if you have a high-rise building that has a fire and you don't really know what's blocked in terms of the stairwells and so forth one of these can fly up the stairwells long as it knows exactly where it's
00:03:09 going to with its own sensors and suddenly you don't have to have the firefighter crew to expose themselves to that risk.