Cambridge – Andrew Barry a PhD student of the MIT has designed a self-flying drone able to detect obstacles and avoid crashes while traveling at 30 miles per hour.

The drone developed in the MIT’s computer Science and Artificial Intelligence lab is armed with an obstacle-detection system that easily allows it to avoid trees, or any other kind of obstacle it could find in its way within a distance of 10 meters ahead.

“You don’t have to know about anything that’s closer or further than that. 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,” Barry explained.

mit-self-navigating-drone
MIT’s Computer Science and Artificial Intelligence Lab (CSAIL) built a drone that can evade objects autonomously. Credit: MIT CSAIL/Newsweek

Apart from avoiding crashes and do it at a considerable speed this automated device can also navigate on its own. The algorithm developed by Barry allows the drone to build a map of its surroundings 20 times faster than any existing software. Normally self-driving drones have maps previously uploaded to help them through the navigation.

This kind of development is quite remarkable, among other things, because a sensor able to do something like this is too heavy to put on a small aircraft. Barry’s drone weights slightly over a pound.

The total cost reached the $1,700. Its design includes two cameras, one in each wing and two computer processors. Because of the weight matters, these processors are as big as the ones found in a cellphone.

The system Barry has developed along with MIT Professor Russ Tedrake is open-source and is now available on line through GitHub. They want to share their algorithm with anyone who wants to give it a try.

Barry hopes to improve his system along with hardware advances as they allow for more complex computation. “This lets us make our algorithms more aggressive, even in environments with larger numbers of obstacles,” the grand student said. He is now working in improving his algorithm to function in different depths. This will permit his drone to navigate through ever crowded surroundings like forests or even a city.

Drones are predicted to do a lot of different tasks in the near future. From picking up targets in war zones to delivering our mail, drones are promising tools to help us improving our quality of life.

Sources: Washington Post