Aerospace engineering sophomore Duncan Miller spent his summer playing with mini quadcopters (helicopters with four rotors) at NASA Langley Research Center. In this video, he demonstrates the "sense-and-avoid" programming he and his fellow interns wrote and tested. That's Miller with the helmet on.
"In the video," Miller says, "I am acting like a human joystick. The quadrotor is matching my helmet's orientation and altitude. When I look left, the quadrotor yaws to the left. When I look to the right, it follows suit. If I get on my knees, it will lower itself to my eye level again."
How does it do this? Tracking cameras around the room have been programmed to recognize certain reflector dots as objects. In the first part of the video, the students had placed the reflector dots on another quadcopter. In the second part, they're on the helmet. The cameras send a signal to a computer, which sends a signal to the quadcopter about what it needs to avoid.
This was actually quite serious work. In addition to his stint in the helmet, Miller focused on hardware and loop simulations and testing decision algorithms. He and his colleagues set up a functioning autonomous vehicle laboratory. Researchers can now use it to study things like swarming autonomous vehicles and aircraft sense-and-avoid issues, says NASA systems engineer Garry Qualls.
The quadcopters the students used are prototype models of the Parrot AR.Drone "flying video game" that hit the market in early September. (The drones don't come with this autonomous mode, though.)