(CAMBRIDGE, Mass.) — This drone doesn’t need help from a remote pilot to find its way around.
The self-flying drone is able to swiftly navigate obstacles while flying through the sky at 30 miles per hour in a new video released by the Massachusetts Institute of Technology’s Computer Science and Artificial Intelligence Lab.
The autonomous drone is so smart it’s able to avoid crashing into people or things, according to Andrew Barry, a doctoral candidate at MIT. Barry developed a stereo-vision algorithm that allows the drone’s software to build a map of the surrounding area in real time, focusing on areas 10 meters away from the drone. The map allows the drone to fly safely through the area without running into anything or needing a human operator.
“You don’t have to know about anything that’s closer or further than that,” Barry told MIT. “As you fly, you push that 10-meter horizon forward, and, as long as your first 10 meters are clear, you can build a full map of the world around you.”
The open-source software operates at 120 frames per second and is able to extract depth information at a speed of 8.3 milliseconds per frame, according to MIT. The drone used in the video has a 34-inch wingspan with cameras on each wing and processors that researchers say are “no fancier than the ones you’d find on a cellphone.”
The technology allows the drone to fly without being weighed down by heavy sensors, Barry said.
“Everyone is building drones these days, but nobody knows how to get them to stop running into things,” Barry said. “Sensors like lidar [a type of radar that uses light] are too heavy to put on small aircraft, and creating maps of the environment in advance isn’t practical. If we want drones that can fly quickly and navigate in the real world, we need better, faster algorithms.”
Copyright © 2015, ABC Radio. All rights reserved.
Recent posts in National News