MIT’s autonomous Drone can evade Obstacles flying at 30mph
Autonomous car, autonomous pods, why not autonomous drone? Yes that is what researcher Andrew Barry, a PhD student at MIT’s Computer Science and Artificial Intelligence have come up with. He launched an autonomous drone fitted with a sophisticated obstacle-detection system that can map its surroundings to successfully evade hindrances, flying at a speed previously unreached by a self-flying UAV.
According to MIT, the UAV can scan obstacles up to 10 meters (32 feet) away at a rate about 20 times faster than current software allows.
Unlike usual drones, the MIT’s drone uses an onboard computer, artificial intelligence, cameras and other equipment to scan and build a real-time map of its surrounding areas. Drone’s camera captures surrounding area at 120 frames per second, reading depth per frame in about 8.3 milliseconds, which enables the drone quickly take necessary movements to evade obstacles.
In a world where drones gradually become part of our daily life like drone deliveries and video shooting, it will have to work its way around many obstacles like buildings, trees, electric pole or even pedestrians. There have been several cases where drones have crashed into things around them. With this new technology in effect, drones should be much smarter thus reduce such accidents.
“Everyone is building drones these days, but nobody knows how to get them to stop running into things,”Barry said. “Sensors like lidar 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.”
Barry hopes to develop more aggressive algorithm which will enable drones scan its surroundings more accurate and faster and even pass through densely populated areas.
“Our current approach results in occasional incorrect estimates known as ‘drift’,”he said. “As hardware advances allow for more complex computation, we will be able to search at multiple depths and therefore check and correct our estimates. This lets us make our algorithms more aggressive, even in environments with larger numbers of obstacles.”
The drone weighs over a pound and has a wingspan of 34 inches. It was built around off-the-shelf components costing around $1,700. And if you are a developers, the good news is that the software is open-source and available on GitHub for download . Take a look at the video below.