Slashdot Mirror
Getting Small UAVs To Imitate Human Pilots Flying Through Dense Forests
New submitter diabolicalrobot writes "The Robotics Institute at CMU has been developing systems to learn from humans. Using a Machine Learning class of techniques called Imitation Learning our group has developed AI software for a small commercially available off-the-shelf ARdrone to autonomously fly through the dense trees for over 3.4 km in experimental runs. We are also developing methods to do longer range planning with such purely vision-guided UAVs. Such technology has a lot of potential impact for surveillance, search and rescue and allowing UAVs to safely share airspace with manned airspace."
To the second moon of Endor.
https://www.youtube.com/watch?feature=player_embedded&v=wfZB0NfMHJM http://www.youtube.com/watch?feature=player_embedded&v=hNsP6-K3Hn4
to the most common aviation accident:
CFIT (human) = Controlled Flight Into Terrain
CFIT (Machine) = Controlled Flight Into Tree
Set your phasers on "funky"!
Violence is just the low hanging fruit of applying robotic vision, well low hanging fruit that pays very well.
There are many applications for this, whether it helps a computer to guide a blind person though a crowded lobby, allowing a driverless car to be safer and more efficient, or even warning you that your car is about to be hit by another car.
Don't know something? Look it up. Still don't know? Then ask.
I never went into robotic vision because nearly all of the immediate applications are military.
Just like radar, and jet engines and rockets and spaceflight and GPS and encryption and antibiotics, yet all those things turned out to be slightly beneficial (or more than slightly, in the case of... all of them) to humanity as a whole.
"None can love freedom heartily, but good men; the rest love not freedom, but license." --John Milton
Link, or Star Wars Version.
When our name is on the back of your car, we're behind you all the way!