airports are now flying robot falcons to chase birds away from your plane
they're called RoBirds
and Canadian airports say mass bird strikes dropped 25% year over year once they started flying them
they were invented because birds flying into jet engines do over $1B of damage to planes every year
> it's what caused the Miracle on the Hudson, when geese took out both engines and the pilot had to land in the river
> so airports try to scare birds off, usually with noise cannons
> but birds learn fast and figure out the bang never actually hurts them, so the cannons stop working
> but the one thing they NEVER stop running from is a peregrine falcon
> so these "biomimetic drones" copy one exactly, with one-meter wings that flap at a falcon's beat and the same paint job
> a pilot on the ground flies them in hunting patterns over the runway
> and after a few rounds the bird flocks decide the whole airport is dangerous and stay tf away lol
nature somehow always has the answers
I think if you'd asked me when I was seven, I would have said that of course by the time I have a kid who's seven, he'll be able to save up his allowance for a few months and buy a flying robot. And here we are.
Reinforcement Learning for Active Perception in Autonomous Navigation.
[📍GitHub & Paper ]
Most robots navigate as if their cameras were nailed in place.
But perception is not passive.
Animals move their heads and eyes constantly to decide where to go next. Robots should do the same.
That is the idea behind “Reinforcement Learning for Active Perception in Autonomous Navigation,” which has just been accepted at ICRA 2026.
Instead of treating navigation and perception as two separate problems, this work trains flying robots to do both at once. The robot does not only decide where to move. It also decides where to look.
Using reinforcement learning, the robot learns to:
•fly safely through cluttered environments,
•actively reorient its onboard camera to reduce uncertainty,
•balance reaching a goal with gathering better visual information.
The key result is that actively controlling perception makes navigation safer.
In simulation and on a real flying robot, this approach consistently outperforms static-camera baselines. The sim-to-real transfer holds up, which is usually where things break.
This is a small but important shift in how we think about autonomy. Better planning does not always come from better maps or bigger models. Sometimes it comes from simply looking in the right direction at the right time.
Thanks for sharing, Kostas Alexis!
📍Code:
Paper:
Video:
——
Weekly robotics and AI insights.
Subscribe free: