We have robots for sweeping floors, mowing lawns, and playing chess. So why not a helicopter that can perform its own flips, rolls, and pirouettes?
Computer scientists at Stanford University on Tuesday said they've developed artificial intelligence software for a helicopter so that it can learn and perform its own aerobatic maneuvers. The hope is that the robotic helicopters could one day be deployed in dangerous missions like putting out wildfires.
The robotic helicopter "watches" another helicopter flown by a human expert, meaning that it records data on its movements, such as position by GPS and velocity. The robot then adapts those maneuvers with new controls every 20th of a second, according to the Stanford computer scientists.
Stanford professor Andrew Ng said that the technology is a demonstration of "apprenticeship learning," in which robots learn by observing an expert, rather than by way of a code.
"In order for us to trust helicopters in...mission-critical applications, it's important that we have very robust, very reliable helicopter controllers that can fly maybe as well as the best human pilots in the world can," Ng said in statement.
Here's Stanford's demo: