Developing autonomous underwater robots for varied applications
Category: research Video duration: Developing autonomous underwater robots for varied applications
Faculty and graduate students are developing teams of underwater robots to autonomously work together, to solve problems primarily in applications in survey and search.
We're at Crater Lake near the Blacksburg campus. And we're working on problems of multiagent coordination for underwater robots. So we're trying to get teams of underwater robots to autonomously work together to solve problems. And primarily applications in survey research, trying to find things, trying to map things. And the only way to do that fast is to have lots of systems working collaboratively. One big step of having a team of robots collaboratively work to perform a search mission is each vehicle has to be able to some portion of that search commission. So today we're working on a single vehicle, fulfilling it's part of what will ultimately be a team mission. We want to see that the hats that it plans look reasonable. That the, the learning of the environment that it's doing looks reasonable. So that when it comes time to integrate into a larger team or send it into a more challenging environment that we can trust it. Ultimately, we want to send these vehicles into the unknown to make decisions, to learn about the environment. Underwater robotics is very cross-disciplinary and our center is unique for our ability to tap into world-class experts in hydrodynamics and marine propulsion, as well as experts and autonomy and control theory. And we also tapped into people who are experts in antennas and all sorts of things that contribute to make an underwater robots work in these very challenging environment. So I started collision avoidance for a Thompson water vehicles. And more specifically, I'm looking at using very limited information from just a few beams from a forward-looking sonar. The issue with the current systems is they're large and expensive. So we're trying to knock those down and just use a few beams instead of the many beams that are currently in use. So we'll do a lot of our initial testing here where it's generally pretty easy to do the work. Then to show that we've really learned how to do it, we'll take it out to the Atlantic Ocean or the Gulf of Mexico or the Chesapeake Bay. There are a lot of problems we're addressing. The main theme is it's really difficult to work in our oceans. We don't have a great sense of what's there. And so we're developing robotic systems that can get into the ocean, looked for things, map things very efficiently. We are working with noaa to find sources of methane in the oceans, were working with a navy to find underwater mines. And to do so very, very fast and a wide variety of applications that are very similar. We test a lot. So to get the systems to work in are very challenging environments. A lot of hands-on effort. It's a large systems integration problem. But at the end of the day, we've got to get our system to work in the field. And that's, that's a big effort.