Blogs

Teaching drones to fly in cities by emulating cars and bikes

Posted by Chris Anderson on January 23, 2018 at 4:46pm

I am happy to share the paper, dataset, code and video "Dronet: Learning to Fly by Driving", where we use Deep Neural Networks to teach drones to navigate autonomously in the streets of a city by imitating cars and bicycles, thus respecting traffic rules! http://rpg.ifi.uzh.ch/dronet.html

E-mail me when people leave their comments –

You need to be a member of diydrones to add comments!

Join diydrones

Comments

andre van calster January 28, 2018 at 1:51am

congrats
Andreas Gazis January 25, 2018 at 10:05am

Well, it's coming from the Robotics and Perception Group of ETH. PX4 is from the Computer Vision and Geometry Group of ETH.

The Bebop can run the original Parrot, PX4 and Arducopter software stacks. My guess is that these guys being ETHians would probably have gone for PX4.

DroNet: Learning to Fly by Driving

DroNet Learning to Fly by Driving
robert bouwens January 25, 2018 at 9:01am

listen to the dialect - züri nord - the eth is close.

but it is a bebop 2 ;-)
Andreas Gazis January 25, 2018 at 8:44am

Holy cow, this is spectacular! This being ETH Zurich, I presume they used PX4?

This reply was deleted.

#xg { position:relative;top:120px; } #xn_bar { top:120px; }

Hello, you need to enable JavaScript to use diydrones.

Please check your browser settings or contact your system administrator.