Unmanned aerial vehicles (UAVs) are vital to other areas of research as well. For example, a team of researchers from the Scripps Institution of Oceanography use UAVs to provide new insights into how air pollution can impact the atmosphere's albedo, or rate of sunlight reflection. Learn more in this video and news release. Credit: Scripps Institution of Oceanography, UC San Diego
The 2008 Summer Olympics in Beijing gave scientists a once-in-a-lifetime opportunity to observe how the atmosphere responds when a heavily populated region substantially curbs everyday industrial emissions. Find out more about the "Cheju ABC Plume-Monsoon Experiment" (CAPMEX) in this news release.
Credit: Scripps Institution of Oceanography
Nereus, a hybrid remotely operated vehicle (HROV), dove to 10,902 meters (6.8 miles) in May 2009, making it the world's deepest-diving vehicle. Learn more in this news release.
Credit: Woods Hole Oceanographic Institution (WHOI)
Providing the benefits of speed, portability and access, a pair of unmanned aerial vehicles (UAVs) surveyed storm-damaged communities in Mississippi as part of the search for trapped survivors of Hurricane Katrina in 2005.
See more in this video and news release.
Credit: Safety Security Rescue Research Center
Researchers at the University of Pennsylvania have found an effective way of getting students interested and excited about science and engineering --by teaching them how to design, build and operate robots.
October 13, 2009
Unmanned Helicopters Could Help Air Traffic Controllers
Small aircraft could have big impact on safety
The chilling video of a small plane and helicopter colliding over the Hudson River in early August showed that technology can't always protect the crowded skies. Nine people died in that crash in New York City.
Some answers to better collision avoidance systems may come from unusual tools: small, autonomous helicopters that are getting smarter all the time.
"We're interested in developing automated collision avoidance algorithms that can be used for civilian aircraft in the air traffic control system," said Claire Tomlin, professor of electrical engineering and computer sciences at the University of California, Berkeley and professor of aeronautics and astronautics at Stanford University.
Starting with test aircraft
The quadrotors developed by Tomlin and her team are about two feet by two feet, snap together like Legos and look more like toys than airliners. But, the technology the quadrotors are used to test could translate to systems that better protect the flying public.
"This is far from a civilian aircraft, but once you take the dynamics of the aircraft and you plug in the dynamics of our quadrotor, the basic algorithms are the same," explained Tomlin.
Tomlin used a CAREER (Faculty Early Career Development) award from the National Science Foundation (NSF) to create the first of these aircraft. She and her graduate students are now on version four.
"I wanted to do something practical and something that had an immediate, practical relevance," said Tomlin.
She first learned about air traffic systems working in the Air Traffic Automation Group at NASA Ames.
"It's a huge, complex system, involving a lot of aircraft, equipment, humans and decision-makers," she said. "And it's a wonderful control problem, when one thinks about how you could use automation to improve the efficiency of the overall system."
No fly zones
In one test with the quadrotors, students who are controlling the aircraft try to get them to crash into each other, but a collision avoidance program based on "reachable set technology" kicks in, and creates a safe "no fly zone."
"As soon as the vehicles detected that they were coming close to each other's no fly zones, the automated algorithm onboard would take over and say, 'No, you can't do that,' and it would safely guide the vehicles along sort of the boundaries of these sets," explained Tomlin.
But to be of the most value to humans, for instance in navigating narrow or dangerous places, the 'bots' must have the tools to make decisions on their own, and adding cameras and other sensors is making that possible.
"The vehicle will look out and sort of see whether or not there are obstacles in its path such as other vehicles, or conversely, whether there’s something interesting that it should get closer to in order to track," said Tomlin.
Tomlin's team is able to show off a little. The vehicles can now safely complete an autonomous back-flip, which is not a simple procedure. Among other things, the motors have to be turned off briefly when the aircraft is upside down.
"We break it into three modes," explains graduate student Jeremy Gillula. "The idea during the impulse phase is we want to control the torque on the motors to start the rotation of the vehicle. Then for the drift phase--that's when the motors are off--it's basically just drifting through, it is continuing the rotation it started. And finally, there is the recovery phase, where the motors come back on, and it essentially catches itself. As it is basically falling and flipping, it rights itself, stabilizes itself, and comes back to a safe hover."
"The back-flip is accomplished without any humans at the controls. The level of human input is we hit the 'go' button. So while the back flip has a great "wow" factor, there's an important technical reason for doing it. The back-flip is nice but it's the fact that there is the safety analysis behind it," Gillula continued.
"You think about robots in general. You think about UAV (unmanned aerial vehicles) drones flying. You want to make sure there are some safety guarantees. Because, not only are they expensive machines, but they are flying above people. So you want to make sure that they behave as you expect them to behave," he added.
These unassuming aircraft, with off the shelf parts, may not look like lifesavers. But they provide a path to building airplanes and helicopters that are smart enough to navigate safely on their own.
"It's a lot of fun to work on," said Tomlin. "Frustrating at times, but then very rewarding at times."