An Indoor Vision Guided Flying Robot

Posted 18 Aug 2005 at 15:31 UTC by steve Share This

Roland Piquepaille writes "It's not always easy to explore small buildings in dangerous areas and even more difficult to see what might be hidden in a cave or a tunnel. In a short article, the MIT Technology Review describes the results obtained by Swiss researchers with a small robotic aircraft. It only weighs 30 grams for a 80-centimeter wingspan and can be flow inside a building for about 4 minutes. With its two 1-gram cameras, a gyroscope, and a small microcontroller onboard, it can detect walls and automatically avoid collisions. The team is now working on even smaller versions of these flying robots which will be used for search-and-rescue, reconnaissance, and inspection applications." For more details see Roland's latest blog entry. The article is based on research presented in the recently published paper, Toward 30-gram Autonomous Indoor Aircraft: Vision-based Obstacle Avoidance and Altitude Control (PDF format) by researchers at the EPFL Autonomous Systems Lab. The robot uses Optical Flow to understand the motion it sees. This is an under-utilized biologically inspired vision processing method that's always been a favorite of mine for robotics applications.

Optical flow., posted 20 Aug 2005 at 00:11 UTC by Pontifier » (Apprentice)

I spent months working on a vision system for my darpa challenge vehicle. In the end I realised that tracking arbitrary (low texture) surfaces is difficult. our eyes can pick out details to lock onto even on aparently featureless surfaces. Camera noise, and motion blur make this much more difficult as well. Looks like they ran into the same problem as they require a high contrast environment, and are working towards a better camera system. I still feel like stealing motion vectors from MPEG encoding might be a good method for finding optical flow.

There are videos at

See more of the latest robot news!

Recent blogs

30 Sep 2017 evilrobots (Observer)
10 Jun 2017 wedesoft (Master)
9 Jun 2017 mwaibel (Master)
25 May 2017 AI4U (Observer)
25 Feb 2017 steve (Master)
16 Aug 2016 Flanneltron (Journeyer)
27 Jun 2016 Petar.Kormushev (Master)
2 May 2016 motters (Master)
10 Sep 2015 svo (Master)
14 Nov 2014 Sergey Popov (Apprentice)
Share this page