Grace Gao receives NSF CAREER award for UAV research


Debra Larsen, Aerospace Engineering

The use of autonomous systems such as self-driving cars and unmanned aerial vehicles (UAVs) is growing. To safely navigate autonomous vehicles, reliable and accurate information from the Global Positioning System is critical. CSL researcher Grace Gao, assistant professor in the Department of Aerospace Engineering at the University of Illinois, was recently funded through the National Science Foundation Faculty Early Career Development (CAREER) Program to conduct research on the subject.

Grace Xingxin Gao
Grace Xingxin Gao
Gao’s research project will study, not just the accuracy of GPS information, but also its integrity. “Integrity, in this case, means that you have a certain confidence level,” Gao said. “You not only want to navigate accurately, but also know the confidence level of the accuracy.” It’s like when someone qualifies their answer to a question by saying that they are 99 percent confident that they are correct. Gao wants to see the same sort of measurement for GPS coordinates for UAVs. 

Gao said although positioning integrity has been well addressed in the Federal Aviation Administration to guide and land commercial aircraft, it has not been well studied for many emerging autonomous navigation positioning integrity such as self-driving cars and small UAVs. And unfortunately, the FAA systems won’t work in many UAV circumstances.

“UAVs, such as drones and small aircraft, autonomous cars, and even regular cars using GPS navigation, often have to navigate in urban environments,” Gao said, “In that kind of environment, GPS signals can be blocked or reflected by buildings. Compare that to an airplane flying in open sky where you have no signal blockage or GPS obstructions. For example, when you’re behind a building and the signal is blocked, you will get a message saying ‘hey, although your GPS can tell you that you are here, we are not very confident in that information.’ If you’re not very confident in your GPS location, you might have to slow down, or hover in place. If you’re in an autonomous car in an urban environment with your hands off the wheel, and you’re not confident about the GPS location you’re getting, you might put your hands back on the wheel and pay more attention to your surroundings.”

There has been a lot of work on trying to make the navigation positioning more accurate, but Gao’s research seeks to find a better way to assess and monitor the confidence level and to quantify the accuracy of that confidence level. “To achieve that, we use multiple sensors. We pair up GPS with camera vision and other sensors. Traditionally, people try to eliminate the signal reflection off of buildings, treating it as an error source. But instead of trying to mitigate or get rid of what we call multipath we want to utilize that signal as an additional navigation signal.”

NSF CAREER proposals are written by individual early-career investigators and include research and education activities that are integrated, innovative, and ambitious.  Gao has already demonstrated innovation in the novel method she developed to test a drone’s accuracy in an urban environment without the danger of injuring pedestrians, buildings, cars, or other structures.

“We didn’t want to simulate the reflection off of buildings,” Gao said. “We wanted real-world data.” To do that, she flew the drone through the streets of U of I’s campus town, while safely tethered inside a cage made of netting attached to a frame, all on the back of a tow truck. Video of Gao’s UAV flight in its “mobile lab” is available online.

Gao said she strives to do research that it is at the connection between academic value and real-world application.