[CMU NREC] FALCO drone: Safe flights by agile collision avoidance

Video shows fast flights of a small drone built by the National Robotics Engineering Center (NREC) of Carnegie Mellon University upon the Cube Orange flight controller, flying at ~3m/s through tight passages and avoiding static / moving obstacles.

No prior information about the environment or visual fiducial is used. The aerial system detects free space by collision avoidance technology using range data from a front-facing depth camera and guides the drone to fly through the free space. State estimation is provided by a rear-facing tracking camera.

All custom software runs on a Raspberry PI 4B computer with an 1.8GHz ARM processor. The collision avoidance and path following modules together consume only ~20% of a single core on the computer.

CMU National Robotics Engineering Center (NREC)
https://www.nrec.ri.cmu.edu

2 Likes

Pushed a PR for example use-case:

old_video:
https://www.youtube.com/watch?v=0QlaItqLTPc

Placeholder for future follow-up info

The github link in the video description is dead. Can you find a replacement?

1 Like

The Aerial Navigation Development Environment will be released in May 2022.
Thank you for asking :pray:

Are you one of the authors?
can you remove the .css changes from the Pull request?

1 Like

Yes.

I just removed the extra .css from the PR (it was auto-generated by the update.py for some reason). Also squashed all commits into one :pray:.