Artificial IntelligenceCryptocurrencyMetaverseCybersecurityTech Forward

Caltech Researchers Are Training a Drone to Herd Flocks of Birds Away from Airports

August 7, 2018, 11:36 PM UTC

Researchers have developed an algorithm that can pilot an unmanned, off-the-shelf drone to herd flocks of birds away from the airspace surrounding airports.

The inspiration for the research came from US Airways Flight 1549, which in 2009 struck a flock of geese shortly after taking off and was safely navigated into a landing on the Hudson River by pilot Chesley “Sully” Sullenberger. The idea is to use drones to avoid such scenarios in the future.

The algorithm, developed by researchers at Caltech, draws on computer models of the behavior of birds inside flocks, such as collision avoidance and velocity matching. It can allow a single drone to fly alongside flocks of birds to adjust their paths away from air traffic.

Current strategies to keep flocks away from airplanes include trained falcons to scare birds off and using hand-piloted drones. Both of those strategies can be costly, and the human-navigated drones to often steer too close to the birds.

“When herding birds away from an airspace, you have to be very careful in how you position your drone,” Soon-Jo Chung, an associate professor at Caltech who led the robotic-herding research, told the science-news site TechXplore. “If it’s too far away, it won’t move the flock. And if it gets too close, you risk scattering the flock and making it completely uncontrollable. That’s difficult to do with a piloted drone.”

The algorithm steers the drone to approach just close enough to a flock to appear to be a threat but without panicking the birds into dispersing into their own individual, unpredictable paths.

Chung’s team tested the algorithm in a remote area of Korea on a flock of egrets and a flock of loons. They discovered that, when the drone approached a flock gradually and then flew alongside at a certain distance, it had the desired effect of herding the birds.

A paper discussing the robotic herding noted that further experiments need to take place before the algorithm is ready to take flight in the wild.