Automatic Re-Initialization and Failure Recovery for Aggressive Flight with a Monocular Vision-Based Quadrotor

We demonstrate a new method to automatically recover and stabilize a drone from any initial condition, without GPS. On the one hand, this new technology will allow drones to be launched by simply tossing them in the air, like a “baseball”. On the other hand, it will make drones much safer since they can recover stable flight after a system failure. This is crucial for commercial use of drones, such as parcel delivery.

Our quadrotor is equipped with a single camera, an inertial measurement unit, and a distance sensor (Teraranger One). Since we use only onboard sensors and no GPS, the method works both indoors and outdoors.

Details about the system can be found on:
http://rpg.ifi.uzh.ch/aggressive_flig…

And in our paper:

M. Faessler, F. Fontana, C. Forster, D. Scaramuzza, Automatic Re-Initialization and Failure Recovery for Aggressive Flight with a Monocular Vision-Based Quadrotor, IEEE International Conference on Robotics and Automation, Seattle, May, 2015
http://rpg.ifi.uzh.ch/docs/ICRA15_Fae…

Video credits:
Flavio Fontana, Matthias Faessler, Christian Forster, Elias Mueggler, and Davide Scaramuzza

Robotics and Perception Group, University of Zurich
http://rpg.ifi.uzh.ch/

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.