Video Friday: Flocking Quadcopters, Hip Hop NAO, and Watson Gets Cooking

Let me apologize in advance for this, because it's self-promotional and boring: I'm going to to be moderating a panel for the Commonwealth Club of California on "The State of Personal Robotics," featuring Pieter Abbeel from UC Berkeley, Melonee Wise from Unbounded Robotics, and Joe Augenbraun, one of the founders of Neato Robotics. It's Monday, March 10, in downtown San Francisco, and you should totally come because it's going to be great and I will probably be really awkward the whole time. Oh, and if you can think of any questions that you want me to ask any of these people, let me know, and I'll put them on the spot. More info here.

Okay, enough of that, time for videos!

Animals have somehow figured out how to navigate and avoid collisions even when there's a whole bunch of them trying to occupy the same small space. Swarming and flocking behaviors are just as important for fleets of quadrotors, especially when they're operating autonomously in an outdoor environment.

According to [Tamás Vicsek, a physicist at Eötvös Loránd University in Budapest], the only other truly autonomous drone flock was created in 2011 by robotics researcher Dario Floreano at the Swiss Federal Institute of Technology in Lausanne. But his machines were fixed-wing fliers that could move only at constant speeds and had to fly at different heights to avoid collisions. “It looked like a swarm but wasn’t a real one because they didn’t interact with one another,” says Vicsek. By contrast, his drones can coordinate their movements to form rotating rings or straight lines. If Vicsek tells them that they face a wall with a gap in it, they can queue up to squeeze through.

The drones did not flock successfully until the team managed to speed up their reaction times — a challenge that Floreano and his team also had to overcome. Vicsek and his team submitted their results as a presentation at the International Conference on Intelligent Robots and Systems, to be held next September in Chicago, Illinois.

That's IROS 2014, and we'll be there to bring you more on this.

[ COLLMOT ] via [ Nature ]

 


 

Late last year, we showed you the ARES robotic cargo carrier concept from Lockheed Martin's Skunk Works, and they've just come out with this animation showing how it's all supposed to work:

[ ARES ]

 


 

Sooo, a giant robot mantis is under development over at CMU:

Because that's not going to be scary at all.

[ Chiara ]

 


 

Drone delivery doesn't work. Netflix gets this, for which we applaud them:

Via [ Hacking Netflix ]

 


 

Robots are tools with which you can find ways to combine all of your passions, if you're clever and put some effort into it. Audrick Fausta, a mechatronics engineer at Aldebaran Robotics, taught his NAO to be a hip hop dance partner:

And here's how he did it:

[ Aldebaran Robotics ]

 


 

This next video shows the Shadow Hand performing grasp stabilization control on various objects. Were I to guess, I would say that "grasp stabilization control" is a fancy roboticists phrase for "not dropping stuff."

[ Shadow Robot Company ]

 


 

Here is another video with a ridiculously complicated premise: "Optimal Parameter Identification for Discrete Mechanical Systems with Application to Flexible Object Manipulation." What I think it's about is using the arm of a Baxter to jiggle a flexible object (a bicycle tire, in this case) in order to create a model of that object so that you can more effectively mess with it later:

I may have gotten that totally wrong, though. Fortunately, the paper is available here.

 


 

Seeing Geminoid DK next to its human never ceases to creep me out a little bit:

Robots for the win, a sentiment we can all support.

[ India Today ]

 


 

The University of Maryland's Robo Raven now has a camera on it, giving a robobird's eye view of some flappy playground dive-bombing:

[ UMD Robotics ]

 


 

This video summarizes FZI's results from the DEXMART project.

"DEXMART" stands for "DEXterous and autonomous dual-arm/hand robotic manipulation with sMART sensory-motor skills: A bridge from natural to artificial cognition". DEXMART was a large-scale integrating project, which was funded under the European Community's 7th Framework Programme.

The project started on the 1st of February, 2008 and ended in 2012.

DEXMART had the ambition to fill the gap between the use of robots in industrial environments and the use of future robots in everyday human and unstructured environments, contributing to domains of personal and service robotics where dexterous and autonomous dual-hand manipulation capabilities are required. FZI contributed methods for learning, adaption and execution of bimanual manipulation tasks. The contributions were presented in a series of demonstrations. Some of these demonstrations are shown in this video.

[ DEXMART ]

 


 

This last vid is more of an artificial intelligence thing than a robot thing, but we're going to forgive that because of how tasty it is:

[ Computational Creativity from IBM and Watson ]

Advertisement

Automaton

IEEE Spectrum's award-winning robotics blog, featuring news, articles, and videos on robots, humanoids, automation, artificial intelligence, and more.
Contact us:  e.guizzo@ieee.org

Editor
Erico Guizzo
New York, N.Y.
Senior Writer
Evan Ackerman
Berkeley, Calif.
 
Contributor
Jason Falconer
Canada
Contributor
Angelica Lim
Tokyo, Japan
 

Newsletter Sign Up

Sign up for the Automaton newsletter and get biweekly updates about robotics, automation, and AI, all delivered directly to your inbox.

Advertisement