New video from ETH Zurich’s Flying Machine Arena, you can read more on Robohub:
http://robohub.org/video-watch-flying-robots-cooperate-throw-and-catch-balls/
PS. Robohub.org is an online platform that brings together leading communicators in robotics research, start-ups, business, and education from around the world. We just launch the beta version, take a look! http://robohub.org/
Comments
There are two halves to the Quad problem. 1) Determining Position / Orientation (along with translational velocity and angular velocity). 2) Control. How to make the Quad accurately move in the desired directions over time.
This video amazing sophisitacation of 2), the second half of the problem .
Alot of people, including ETH, are working on 1), in separate projects.
@Chris: it was really nice to meet you! Hope you enjoyed the demos!
@Andrew thanks for the kind(er) words. It's amazing what people write online. I guess also using real names really helps with the level of comments.
The 007 theme video was done by the guys at GRASP lab, UPenn. These vehicles can fly outside just fine -- obviously not with the same precision/convenience as in a dedicated mocap space.
@crashpilot, etc. These are extremely harsh (to a meaningless level) comments. Good luck with that.
I was there this week for this! Yes, it's a Vicon motion capture space. Here's a video I took on my cellphone as they were tying to get the system to recognize the ball:
@Marooned: I know, but this populistic exposition, like all their presentations (007 Theme etc) - i just can't see anymore.
Experiments must reduce the environment factors to properly measure a variable. ETH Zurich is using vicon to simplify their research problem. Many of these researchers are also testing outdoor experiments and soon we will see the transfer of research lab to outdoor work. Please don't judge without carefully reading their research papers!
Crashpilot: I must disagree. With your thinking we would be still in stone age. Most inventions had its beginnings in labs under certain conditions. Then after proper analysis scientists could lower restrictions on conditions and eventually the idea became a real world stuff.
Some scientists playing around in an artificial world to justify their payment. Good news: There is a real world outside your lab! Go there and do the same thing without cameras in every tree, with sunstorms disturbing your mag, GPS etc with limited visibility etc - that would impress me. This simply sucks completely.
Pretty sure they model prediction algorithm.
WoW! The quads are sooooooooooooooooooooooooo stable, its amazing!
Very neat video.
Francisc: FMA has their own custom controllers. A closeup from their website shows the controller.
-
1
-
2
of 2 Next