Video Friday is your weekly selection of awesome robotics videos, collected by your shapeshifting Automaton bloggers. We’ll also be posting a weekly calendar of upcoming robotics events for the next few months; here’s what we have so far (send us your events!):
RoboCup European Open – March 30-4, 2016 – Eindhoven, Netherlands
WeRobot 2016 – April 1-2, 2016 – Miami, Fla., USA
National Robotics Week – April 2-10, 2016 – United States
AISB HRI Symposium – April 5-6, 2016 – Sheffield, United Kingdom
ROS-Industrial Training Class – April 6-8, 2016 – San Antonio, Texas, USA
Robotics in Education 2016 – April 14-15, 2016 – Vienna, Austria
NASA Swarmathon – April 18-22, 2016 – NASA KSC, Fla., USA
LEO Robotics Congress – April 21, 2016 – Eindhoven, Netherlands
International Collaborative Robots Workshop – May 3-4, 2016 – Boston, Mass., USA
ICARSC 2016 – May 4-6, 2016 – Bragança, Portugal
Robotica 2016 – May 4-8, 2016 – Bragança, Portugal
ARMS 2016 – May 9-13, 2016 – Singapore
ICRA 2016 – May 16-21, 2016 – Stockholm, Sweden
NASA Robotic Mining Competition – May 18-20, 2016 – NASA KSC, Fla., USA
Skolkovo Robotics Conference – May 20, 2016 – Skolkovo, Russia
Innorobo 2016 – May 24-26, 2016 – Paris, France
RoboCity16 – May 26-27, 2016 – Madrid, Spain
RoboBusiness Europe – June 1-3, 2016 – Odense, Denmark
IEEE RAS MRSSS 2016 – June 6-10, 2016 – Singapore
CR-HRI – June 6-10, 2016 – Orlando, Fla., USA
NASA SRRC Level 1 – June 6-11, 2016 – Worcester, Mass., USA
Field Robot Event – June 14-18, 2016 – Haßfurt, Germany
RSS 2016 – June 18-22, 2016 – Ann Arbor, Mich., USA
European Land Robot Trial – June 20-24, 2016 – Eggendorf, Austria
Automatica 2016 – June 21-25, 2016 – Munich, Germany
ISR 2016 – June 21-22, 2016 – Munich, Germany
Let us know if you have suggestions for next week, and enjoy today’s videos.
HEBI Robotics, who we like because they make snake robots that also work as legs for non-snake robots, have just released a new modular robot actuator that’s geared towards general-purpose robotics and looks pretty awesome:
This series of powerful robot module allow engineers, researchers, and industrial integrators to quickly and easily create world-class custom robots of any configuration. These actuators are packed with sensors that enable controllable position, velocity, and sensitive torque control as well as three axis inertial measurement.
Each module is a series-elastic actuator that integrates a brushless motor, geartrain, spring, encoders, and control electronics into a compact package that runs on anything from 18V-50V DC and communicates using standard 10/100Mbps Ethernet. This module is designed to function as a full-featured robotic component as opposed to a simple servo motor.
The output rotates continuously, requires no calibration or homing on boot-up, and contains a thru-bore for easy daisy-chaining and wiring. This enables these modules to be used in everything from wheeled robots to multi degree of freedom collaborative robotic arms.
And what can you do with these actuators? You can use them to build a 6-DoF articulated arm in just 24 minutes. Or a SCARA-type robot to help you make . . . more robots. Robots building robots!
[ HEBI Robotics ]
UIUC’s bat robot is getting more and more bat-like, and can now independently perform closed-loop indoor flight:
We’ll get more details on this at ICRA 2016 in May.
[ UIUC ]
Robots hate hate hate hate hate stairs. They hate them. Stairs are the worst. We’ve seen all kinds of creative attempts to reliably get robots up stairs, but Transcend Robotics has come up with a platform that does it effortlessly:
What makes this robot unique is that it doesn’t require the deployment and control of independent tracked arms in order to tackle stairs: with ARTI, you just drive right at the stairs (or any other tall obstacle) and and go straight up and over without stopping. The hardware is simpler and less expensive, and the robot is much easier for a user (especially an inexperienced user) to control.
Jiri Zemanek used an EggBot to draw these patterns, which animate when the egg is rotated at a specific speed in front of a strobe light or camera:
“Various patterns are generated in Matlab using mathematical equations similar to ones describing Spirograph (or harmonograph) and Phyllotaxis. The patterns are calculated in such a way that when rotated under a stroboscopic light of suitable frequency or when recorded by a camera, they start to animate. It is kind of zoetrope---early device for animation. Eggs were painted using EggBot (designed by Bruce Shapiro as open hardware and available as a kit from http://www.evilmadscientist.com/). To draw on eggs, we used standard permanent markers and an electro kistka with bee wax followed by dying. Eggs are rotated at a constant speed, special for each pattern, by a brushless motor. No computer graphics tricks are used in the video.”
[ Eggstatic ]
Jibo will help you order you a pizza and is considerate enough to try and prevent you from killing your friends while doing so:
[ Jibo ]
Kuka. Listen. You have good robots. Let them play, for real, no tricks. We promise we’ll be impressed, even if the robots lose, but please, no more overhyped underperformances, okay?
[ Kuka ]
Here’s a cool demo showing what you can do with a six-axis force-torque sensor on a Robotiq gripper:
It’s the Robotics Innovation Facility at the Bristol Robotics Laboratory, where they have a facility to do innovation in a laboratory of robotics! In Bristol!
[ RIF@Bristol ]
If you don’t want to wait for an Alpha 2 robot, the Alpha 1S looks decent, and this video seems to show it doing mostly realistic stuff. It’s only $450 and available on Amazon.
RoboProTip: high kicks are easier without pants on. And this is why robots don’t wear pants.
[ UBTECH ]
Is your interface precise enough to literally thread a needle with an industrial robot? UT Austin can do that:
Panasonic understands that in Japan, robots are going to be a critical part of the future, in many different areas:
ESA’s vision for a moon village includes a lot of robots, of course:
[ ESA ]
We’re not saying that this is the beginning of the Terminator T-1000, but it’s totally the beginning of the Terminator T-1000:
“That’s what this idea is about, to have a skeleton when you need it, melt it away when you don’t, and then reform it,” [Cornell engineering professor Rob Shepherd] said.
Shepherd said this material would be the skin for a morphing wing, giving the MAV the ability to become an underwater vehicle on the fly. “If you have a wing that’s really broad, you can’t do that because the wing will break off when it hits the water,” he said. “So you need to sweep it back, similar to what a puffin does, and then go under water. And using that new shape, it could be a propeller-driven ship.”
In addition to a morphing-wing application, Ilse Van Meerbeek, a graduate student in the field of mechanical engineering and the first author of the paper, sees this material being used in soft robots that must negotiate tight spaces. “It could be used in search-and-rescue robots,” she said. “It would be able to go into dangerous and/or unpredictable environments, and be able to go through narrow cracks, which rigid robots can’t do.”
[ Cornell ]
Episode 2 of “Moon Shot,” a documentary on the Google Lunar XPRIZE:
“Founded by an ad hoc group of part-timers, this Berlin-based GLXP team plans to open source its mission data. Team leader Robert Böhme, who was raised in the former East Germany, says the free exchange of information is ultimately more important than money.”
Finally, CMU RI Seminar: Neville Hogan, from MIT.
Human dexterity and agility vastly exceed that of contemporary robots. Yet humans have vastly slower ‘hardware’ (e.g. muscles) and ‘wetware’ (e.g. neurons). How can this paradox be resolved? Slow actuators and long communication delays require predictive control based on some form of internal model—but what form? I will argue that a plausible answer is based on dynamic primitives; they enable highly dynamic behavior with minimal high-level supervision and intervention. Controlling physical interaction requires mechanical impedance to be among the classes of dynamic primitives. I will review how pre-computing appropriate mechanical impedance may be cast as an optimization problem, provided the objective function includes both force and motion at an interaction port. To combine both motion and interaction primitives, I propose a nonlinear generalization of the classical equivalent circuit. It reconciles contrasting constraints of information-processing (computation) and energy-processing (physical dynamics). I suggest that nonlinear equivalent networks provide a general basis for the internal models required for high-performance interactive control.
[ CMU RI Seminar ]