Back in April, we posted about a micro UAV from Stanford University that uses spines to land and perch on vertical surfaces. In order to do this, the UAV has to sense its proximity to the surface and then execute a well timed pitch up and stall maneuver. This approach is fairly complex, at least when you compare it to the system under development at EPFL, which basically involves the UAV just smashing headfirst into whatever it wants to perch on.

Mirko Kovac, the same guy who came up with the jumping locust robot, attached some very sharp needles to the nose of a micro UAV. The needles deploy on impact, and are capable of sticking the aircraft to wood and even concrete. To detach, a small motor withdraws the needles, and the MAV detaches and takes flight. The attachment technique is about as simple as it gets, although arguably, this is outweighed by the reliance on the motor to detach, especially since the motor is dead weight the rest of the time. On the other hand, if you don’t get the perch right, the efficiency of your detach technique doesn’t matter in the least.

The big advantage of a perching robot is, of course, that you get all the mobility of a flying platform along with the endurance of a stationary platform. People working on these things talk about the robots being used to help people during natural disasters, which is nice and all, but the subtext is definitely about surveillance. Besides, that’s where the funding is. And there’s not anything wrong with that, it’s just the way it is.

Kovac plans to combine this perching technology with his flying jumping robot to make an awesome flying, jumping, perching robot of awesomeness.

[ EPFL ] VIA [ Physorg ]

My only comment on this shirt is that at $14.95, it’s overpriced by about $20.

And since I can’t imagine any other conceivable opportunity which would allow me to post this, after the jump, you can watch Sir Mixalot’s Baby Got Back music video. Just a warning… This video may not be safe for viewing at work, or for viewing anywhere else if you have taste, as it contains (among other things) pieces of fruit shaped like butts. (Read more…)

When I originally heard about PR2 being taught to push a cart, my comment was “meh,” at least in contrast to PR2 fetching beer or playing pool. However, nobody told me that the robot would be picking up trash, putting it on the aforementioned cart, and then pushing the cart over to the sink. We’re about to go full circle here, folks… PR2 fetches beer. PR2 plays pool while you drink beer. PR2 cleans up beer. Repeat.

Pushing a cart is also trickier than it seems. The physical act of pushing is not particularly challenging, but the cart makes it significantly more difficult for the robot to navigate. Not only does the cart block the PR2′s forward view, but in effect, it changes the shape of the robot, changing how PR2 has to move, especially in tight spaces. The solution to this involved just applying different motion planning software, which was easy to do, thanks to the open source environment of ROS and the standardized PR2 platform.

What I don’t entirely understand is the need for a human in the loop to identify specific containers for cleaning. I guess it might be bad for PR2 to pick up things containing liquid in case it spills, but remember that nifty squeezing experiment from last September that showed how PR2 could tell the difference between open and closed and full and empty containers? Seems like this would be the perfect time to use it.

[ Willow Garage ]

This year’s RoboCup 2010 KidSize championship match was all German, with the FUmanoids from Berlin against the Darmstadt Dribblers. This is no surprise, since it was the exact same final match in 2009. The Dribblers took home the trophy (not sure if there’s actually a trophy), and before you leave unimpressed, remember that this is a fully autonomous match, with the robots relying entirely on their programming and a webcam on their head to find the ball and then decide what to do with it.

More vids, after the jump. (Read more…)

Despite several months worth of pestering, Neato Robotics wouldn’t give us a ship date on the XV-11 robot vacuum that was any more specific than “Summer 2010,” to the point where (if you’re jaded like me) it sort of seemed like we’d never actually see a commercially available product, despite the pre-orders and demo units and stuff. But, Neato has certainly held up their end of the bargain, and if you took the plunge and pre-ordered an XV-11, Neato says it’ll be shipping this week. Yay!

MOUNTAIN VIEW, Calif., June 28 /PRNewswire/ — Neato Robotics™, a Silicon Valley startup pioneering new products to free people from household chores, today announced it will begin shipping its first Neato XV-11™ home robot vacuum cleaners to consumers beginning this week. Based on advanced technologies, Neato Robotics engineers have created a simply smarter robot that thoroughly cleans an area without you needing to be home. Customers who pre-order the highly-anticipated Neato XV-11 at neatorobotics.com will begin receiving them in early July, and additional units will be available for purchase via retail and online partners beginning in mid-July.

The suggested manufacturer’s retail price is $399. The Neato XV-11 includes a 30-day money back guarantee and a one-year limited hardware warranty. In addition to ordering on the Neato Robotics website, the Neato XV-11 will be available for pre-order via retail and online channel partners such as Amazon.com, Hammacher Schlemmer, SharperImage.com and Robotshop.com on July 15 with shipments of new customer orders beginning in August 2010. Additional channel partners will be announced soon.

Also, for those of you who like numbers, this was included as a footnote on the press release:

Based on internal testing, the Neato XV-11′s blower has a flow of 46 cubic feet of air per minute (or 1.3 cubic meters of airflow per minute), which is approximately five times stronger than the closest competitor. It has a vacuum suction of 100 mm of water (mmAq), which is approximately four times stronger than the closest competitor. The combination of this strong suction power with the linear speed of the Neato XV-11 at 24 cm/s creates unprecedented cleaning performance for a robotic vacuum cleaner.

Read our review of the Neato XV-11 here, and our comparison with the iRobot Roomba 560 here.

[ Press Release ]
[ Neato Robotics ]

Pay attention, England: This robot has grit. This robot has determination. This robot is not getting paid anything, and is just playing for the love of the game (or maybe because it’s programmed to). But most importantly, this robot takes a dive at 30 seconds in and then gets right back up again.

Sorry, I couldn’t resist.

This robot is DARwin, from Virginia Tech’s Robotics and Mechanisms Laboratory (RoMeLa). More specifically, this is DARwin-LC (where the LC stands for “low cost”) which made its debut at RoboCup 2010 in Singapore last week. There’s also a DARwin-HP, where HP means “high performance,” implying more expensive servos and sensors. The basic idea is to make DARwin-LC cheap (on the humanoid robot scale), with DARwin-HP out there as an option if you need the extra power and precision.

DARwin is being developed on a National Science Foundation grant specifically to make robotics research, education, and outreach more accessible (read: affordable). As you might expect, the grant goes on to say:

The research results from the humanoid-robotics community will find applications in assistive robotics technology for the elderly, intelligent humanoids, and homeland security and the military.

Today, they play soccer. Tomorrow, they learn how to do community service and (sometimes) be destructive.

Yep, just like professional soccer players. Wonderful.

[ RoMeLa ] VIA [ Robots Dreams ]

Hi, are you drowning? If so, you’re probably anxious to meet EMILY, a robot lifeguard that sacrifices skimpy bathing suits and implanted flotation devices for a sonar system and Jet Ski engine. EMILY (as you’ve probably guessed by now) is an acronym, which doesn’t really stand for EMergency Integrated Lifesaving lanYard but they’re just going to go ahead and make it happen anyway. Seriously, when you start pulling letters out of the middle of words, that’s just not fair.

Moving on.

EMILY is designed to be launched right from the beach or tossed out of a boat or helicopter. Sonar on the robot will somehow identify swimmers making underwater movements that indicate distress and home in on them, or remote control is an option too. Once there, lifeguards (human ones) can communicate with the hopefully conscious swimmer via speakers and cameras and instruct them to grab on to EMILY while the robot tows them to safety.

EMILY’s top speed (unburdened by bedraggled hangers-on) is 28 mph, which is some six times faster than a human. Her endurance of 8+ hours on patrol is nothing to sneeze at either. Basically, her stats make human lifeguards look like slouches. But until EMILY can plant a big wet one and breathe life back into your soggy corpse (or even grab on to a swimmer by herself) she’s not likely to be replacing the cast of Baywatch anytime soon. And let’s be honest: no robot could possibly replace the Hoff.

EMILY costs only $3500, which seems quite affordable. The question is, though, whether EMILY can compete with (say) a human on a real Jet Ski, which remains a much more versatile option than a robot.

[ Hydronalix ] VIA [ PopSci ]

…At least, as long as PR2 doesn’t pick the Bud Light.

[ Willow Garage ]

This little guy may be from 2008, but he’s still pretty cute, and quite talented at going up stairs. This could very well be the fastest stair climbing robot I’ve ever seen. Speed could be a weakness, though, since it looks like the robot needs a pretty significant head start to get up a forgivingly shallow flight of stairs. Going down is no problem, either… The bot just drives straight off, extends its wheels to the ground, and then lowers itself down:

So, pretty cool? Yes. Useful? Well… It’s pretty specialized, and seems to be dependent on getting a running start, but this would be a good bot to have in the lead when we decide to invade M. C. Escher’s House of Stairs, which we totally need to, because wentelteefjes are freakin’ creepy.

[ Chiba ] VIA [ Make ]

Last year, it was a Billie Jean tribute dance. This year, twenty Aldebaran Robotics Naos are kicking it up a notch by dancing in sync to nearly ten minutes worth of music including Maurice Ravel’s Bolero. Never mind the choreography that must have been involved… Consider what ten minutes with no screw ups implies about the precision and repeatability of Nao’s software. Impressive, but until Nao figures out how to do a head spin, I think Manoi can still bust better moves.

VIA [ Wired ]