Video Friday: New Quadruped, Drone UFO, and Bricklaying Robot

Even with two giant robotics events coming up, there's still tons of stuff happening in robotics this week

Image: IIT
Advertisement

It’s shaping up to be a very exciting few weeks in robotics. At the end of this month is ICRA, the IEEE International Conference on Robotics and Automation, from May 26 to 30 in Seattle, Wash. Right after that, the very next week in fact, comes the DARPA Robotics Challenge Finals, June 5 and 6 in Pomona, Calif. We’re going to be at both events, of course, and we’ll be super busy trying to bring you news of everything that’s going on. It’d be great if this week and next week, nothing was happening in robotics so that we could rest up and get all prepared and stuff, but nope: there’s just as much cool new stuff as usual, starting with a brand new hydraulic quadruped robot from the Italian Institute of Technology (IIT).

This is just a teaser, mind you:

Consider yourself teased. We’ve heard that HyQ2Max (this is the next generation of IIT’s HyQ quadruped) is going to make a full appearance within the next few weeks, and we’ll have a full article when it does.

[ IIT ]


It’s the final stretch before the DRC Finals, and teams have (we assume) been working really really hard. Most of them are keeping quiet about what they’re up to, but a few of the braver ones have been posting to YouTube. Here’s IHMC’s ATLAS, doing an admirable job of not falling over, mostly:

Well, it’s “experimental” push recovery, after all.

[ IHMC ]

And here’s the latest from MIT:

“This is a simple example of our approach to perception for the DRC. Faced with a mess of point cloud data, the human provides a few clicks in the user interface to say, e.g., "there is a 2x4 board approximately here" -- these seed a perception algorithm which finds the best local match to the point cloud data.”

Seems like a very efficient way to combine human control with robot perception, especially in a situation where the amount of data that the humans get might be limited.

[ MIT ]


And if you don’t have a million dollar ATLAS to compete in the real DRC Finals, why not hosting your own Mini DRC? That’s what a student at the University of Nevada, Las Vegas, did with his small humanoid Darwin-OP (ROBOTIS recently renamed it just OP, and it’ll unveil a new version, the OP2, soon). We think this is a great idea, and DARPA should officially sponsor a Mini DRC in the future.  

[ UNLV Drones and Autonomous Systems Lab ]

Thanks Jinux!


Hinamitetu, apparently having decided that making little gymnast robots that do horizontal bar is no longer difficult enough, has come up with a new design that can do floor gymnastics, which involves jumping (!) from a standing position:

[ YouTube ]


We’ll be seeing more about this at ICRA, but this video is so cool that it’s definitely worth posting: it’s a soft, wearable robotic glove for hand rehabilitation that allows users to grasp things that they wouldn’t be able to otherwise:

“Soft Robotic Glove for Hand Rehabilitation and Task Specific Training,” by Polygerinos, P., Galloway, K., Savage, E., Herman, M., O’Donnell, K., Walsh, C., will be presented at ICRA in a couple weeks.

[ Harvard Biodesign Lab ]


Aaron Becker does a lot of fascinating research on robotic systems that can operate inside MRI machines, which is a challenging environment (to put it mildly) for anything with metal in it. We’ll have more on his current research at ICRA, but this is a fun video of a strandbeest walking around inside of an active MRI machine thanks to energy harvested from a spinning magnetic rotor:

It was designed by Remy Kaldawy, Aaron T. Becker, Ouajdi Felfoul, and Pierre E. Dupont at Boston Children’s Hospital and Harvard Medical School.


Hey look, Crabster CR200 is doing something practical (finally) by using its underwater manipulation skills to pick up priceless pieces of pottery:

[ KIOST ]


This isn’t a catch-and-release robot for jellyfish, but it is a catch-and-not-shred robot for jellyfish, which is a little better for the jellyfish (I guess?) and way better for people who don’t like getting stung by disembodied jellyfish fragments.

Overpopulated jellyfish has been inflicting enormous damage to marine-related industries. In an effort to minimize this damage, some researchers have proposed jellyfish removal systems including jellyfish shredding or cutting device. However, the removal system with jellyfish shredding device has risks of secondary damages such as eutrophication and sting by fragments of jellyfish if they are venomous. To resolve these risks, a system that removes jellyfish without shredding jellyfish is needed. Thus, we propose a jellyfish removal robot system consisting of an unmanned surface vehicle (USV) and a conveyor device for jellyfish lifting that is attached underneath the USV. The conveyor device consists of an asymmetric funnel-shaped net for guiding jellyfish to the conveyor, a lattice-shaped conveyor belt, and a jellyfish storage. It is also designed to minimize the dimension of the conveyor belt and also minimize the drag force caused by water. The feasibility of the conveyor device was verified at Masan Bay in the Southern coast of South Korea.

“Development of Conveyor-Type Jellyfish Removal Robot System for Venomous Jellyfish,” by Donghoon Kim, Sungwook Jung, Hanguen Kim, Jae-Uk Shin, Taekjun Oh, and Hyun Myung, was presented at the 2015 ICROS Annual Conference in Daejeon, Korea.

Also, these are sea jellies, not jellyfish. Just pointing that out.

[ KAIST ]


One more aquatic robot video, and this one brings you beer! Or, more accurately, it brings you whatever PBR is:

[ OpenROV ] via [ Gizmodo ]


Got a drone? Got nothing better to do? Rig it up with lights and use it to freak out your neighbors! Seriously, that’s what the video description says to do:

[ YouTube ]


Kuka had a massive booth at Hannover Fair this year, and here’s a tour:

[ Kuka ]


Bricklaying strikes me as a task that’s eminently suitable for robots, especially if you want your walls to look all kinds of fancy:

[ Construction Robotics ]


This is certainly not the first robot that can draw, but I like Autodesk’s emphasis on adding in human-like behaviors:

[ Forbes ]


Curiosity has been on the what is literally the longest road trip ever, and here’s an update on the mission from JPL:

[ JPL ]


To finish the week, we have a few videos from WeRobot 2015, a conference on robotics, law, and policy that was held at the University of Washington in April. We weren’t there ourselves, because we’re terrible people, but it looks like all of the panels and talks were recorded. Rather than post 10 hours or so of content all at once, we’ll sprinkle a few in every week for the next couple weeks to let you absorb them gradually.

This first one features Professor Ken Goldberg from UC Berkeley discussing a paper by Dr. Kate Darling from MIT (and also occasionally IEEE Spectrum) on anthropomorphizing robots:

As we increasingly create spaces where robotic technology interacts with humans, our tendency to project lifelike qualities onto robots raises questions around use and policy. Based on a human-robot-interaction experiment conducted in our lab, this paper explores the effects of anthropomorphic framing in the introduction of robotic technology. It discusses concerns about anthropomorphism in certain contexts, but argues that there are also cases where encouraging anthropomorphism is desirable. Because people respond to framing, framing could serve as a tool to separate these cases.

And the second (and last) vid is the WeRobot keynote, in which UW Law professor Ryan Calo talks about robots with Tony Dyson, who has been in robotics for decades but is almost certainly best known as the guy who created Star Wars’ R2D2.

[ WeRobot 2015 ]

Robotics News

Biweekly newsletter on advances and news in robotics, automation, control systems, interviews with leading roboticists, and more.

About the Automaton blog

IEEE Spectrum’s award-winning robotics blog, featuring news, articles, and videos on robots, humanoids, automation, artificial intelligence, and more.