The insane epicness of this movie cannot possibly be overstated. And somehow, the fact that it’s overdubbed in Russian makes it just that much more awesome. It’s called "Enthiran," or "Robot," and if you get a copy, send it to me. Immediately.
Update: the full movie is here on YouTube (all 2+ hours of it), with English captions!
Remember how RoboCup keeps saying that they’ll have a team of fully autonomous humanoids ready to take on BrazilItaly Spain (or whoever) by 2050? RoMeLa’s CHARLI could be the progenitor of those eventual champions, as it’ll be competing in the humanoid adult class at RoboCup 2011. Adult class means that CHARLI is more or less the size of a person; it’s 4 foot 7 which I think is slightly taller than Diego Maradona.
Also heading to RoboCup 2011 from RoMeLa is Team DARwIn and their (slightly smaller) autonomous humanoids:
Not bad, I’d say, and I especially liked those two what I’m going to assume were deliberate fake-outs at the end there. You had me fooled, you tricky little robot you!
Harvest is not your typical robotics start up. For a start, the experience of the founding and management team is unusual for any start up, and truly exceptional in the very young industry of autonomous robotics. Joe Jones alone has more than 24 years of robotics experience. As the first employee of iRobot, he invented the Roomba vacuum cleaning robot, which to date has sold in more than 3 million units, and next to numerous research articles, Jones has also authored three books on robotics, and holds 15 patents. Also, unlike most start ups, the Harvest team had no particular product or application in mind when they first got together in 2007 under the initial name Q-Robotics. Instead, they went on an extended fact-finding mission about possible markets for a robot "one step beyond the Roomba".
What they've come up with is surprising. Agriculture may not seem like the easiest way to start: it's an outdoor application which means dealing with rain, mud, and temperature variations, environments are typically very unstructured, your typical user is unlikely to be highly skilled in software or automation, and tasks are normally accomplished using big machinery - all very different from the Roomba. However, Harvest has found a niche application which eliminates at least two of these challenges and which - they hope - can act as a stepping stone.
Harvest is targeting wholesale shrub farms as the one in the image above, which can have as many as 3 million containers of ornamental plants in 500 acres. Their robots lift and move the plants to perform tasks like spacing potted plants in grids - an operation that has to be repeated as plants grow, to efficiently use space at first, and then again to avoid that they grow into each other. The current human workers will essentially serve as supervisors to Harvest’s robots - they still give direction at a high level, but the machines are doing the heavy lifting. First trials have proven very successful, and the team is now working on a production version.
The Robots podcast interview with Joe Jones gives a deep insight into what it takes to bridge the gap between academia and industry. Jones explains more about the how's and why's of Harvest Automation, why moving potted plants is the next logical step after the Roomba and how they plan to tap into a potentially huge market for autonomous robots.
The DLR hand has the shape and size of a human hand, with five articulated fingers powered by a web of 38 tendons, each connected to an individual motor on the forearm.
The main capability that makes the DLR hand different from other robot hands is that it can control its stiffness. The motors can tension the tendons, allowing the hand to absorb violent shocks. In one test, the researchers hit the hand with a baseball bat—a 66 G impact. The hand survived.
The video below shows the fingers moving and the hand getting hit by a hammer and a metal bar:
The DLR team didn’t want to build an anatomically correct copy of a human hand, as other teams have. They wanted a hand that can perform like a human hand both in terms of dexterity and resilience.
The hand has a total of 19 degrees of freedom, or only one less than the real thing, and it can move the fingers independently to grasp varied objects. The fingers can exert a force of up to 30 newtons at the fingertips, which makes this hand also one of the strongest ever built.
Another key element in the DLR design is a spring mechanism connected to each tendon. These springs [photo left] give the tendons, which are made from a super strong synthetic fiber called Dyneema, more elasticity, allowing the fingers to absorb and release energy, like our own hands do. This capability is key for achieving robustness and for mimicking the kinematic, dynamic, and force properties of the human hand.
During normal operation, the finger joints can turn at about 500 degrees per second. By tensioning the springs, and then releasing their energy to produce extra torque, the joint speed can reach 2000 degrees per second. This means that this robot hand can do something few others, if any, can: snap its fingers.
Why build such a super strong hand?
Markus Grebenstein, the hand's lead designer, says that existing robot hands built with rigid parts, despite their Terminator-tough looks, are relatively fragile. Even small collisions, with forces of a few tens of newtons, can dislodge joints and tear fingers apart.
“If every time a robot bumps its hand, the hand gets damaged, we’ll have a big problem deploying service robots in the real world,” Grebenstein says.
To change its stiffness, the DLR hand uses an approach known as antagonistic actuation. The joints of each finger [photo below] are driven by two tendons, each attached to one motor. When the motors turn in the same direction, the joint moves; when they turn in opposite directions, the joint stiffens.
Other hands, such as the Shadow hand designed in the U.K., also use antagonistic actuation. But the Shadow uses pneumatic artificial muscles, which have limitations in how much they can vary their stiffness.
Before developing the new hand, Grebenstein designed the hand of another advanced robot, the humanoid Justin. He says that in one experiment they would throw heavy balls and have Justin try to catch them. “The impact would strain the joints beyond their limits and kill the fingers,” he says.
The new hand can catch a ball thrown from several meters away. The actuation and spring mechanisms are capable of absorbing the kinetic energy without structural damages.
But the hand can’t always be in a stiff mode. To do manipulation tasks that require accuracy, it’s better to have a hand with low stiffness. By adjusting the tendon motors, the DLR hand can do just that.
To operate the hand, the researchers use special sensor gloves or simply send grasping commands. The control system is based on monitoring the joint angles. It doesn’t need to do impedance control, Grebenstein says, because the hand has compliance within the mechanics.
To detect whether an object is soft and must be handled more gently, the hand measures force by keeping track of the elongation of the spring mechanisms.
“In terms of grasping and dexterity, we’re quite close to the human hand,” he says, adding that the new hand is “miles ahead” of Justin’s hands.
About 13 people have worked on the hand, and Grebenstein insists it’s hard to estimate the cost of the project. But he says that the hardware for one hand would cost between 70,000 and 100,000 euros.
The researchers are now building a complete two-arm torso called the DLR Hand Arm System. Their plan is to study innovative grasping and manipulation strategies, including bimanual manipulations.
Grebenstein hopes that their new approach to hand design will help advance the field of service robots. He says that current robot hardware has limited new developments, because it's costly and researchers can't afford to do experiments that might damage them.
“The problem is," he says, "you can’t learn without experimenting.”
In the first “Matrix” movie, there’s a scene where Neo points to a helicopter on a rooftop and asks Trinity, “Can you fly that thing?” Her answer: “Not yet.” Then she gets a “pilot program” uploaded to her brain and they fly away.
For us humans, with our non-upgradeable, offline meat brains, the possibility of acquiring new skills by connecting our heads to a computer network is still science fiction. Not so for robots.
Several research groups are exploring the idea of robots that rely on cloud-computing infrastructure to access vast amounts of processing power and data. This approach, which some are calling "cloud robotics," would allow robots to offload compute-intensive tasks like image processing and voice recognition and even download new skills instantly, Matrix-style.
Imagine a robot that finds an object that it's never seen or used before—say, a plastic cup. The robot could simply send an image of the cup to the cloud and receive back the object’s name, a 3-D model, and instructions on how to use it, says James Kuffner, a professor at Carnegie Mellon currently working at Google.
Kuffner described the possibilities of cloud robotics at the IEEE International Conference on Humanoid Robots, in Nashville, Tenn., this past December. Embracing the cloud could make robots “lighter, cheaper, and smarter,” he said in his talk, which created much buzz among attendees.
For conventional robots, every task—moving a foot, grasping an object, recognizing a face—requires a significant amount of processing and preprogrammed information. As a result, sophisticated systems like humanoid robots need to carry powerful computers and large batteries to power them.
According to Kuffner, cloud-enabled robots could offload CPU-heavy tasks to remote servers, relying on smaller and less power-hungry onboard computers. Even more promising, the robots could turn to cloud-based services to expand their capabilities.
As an example, he mentioned the Google service known as Google Goggles. You snap a picture of a painting at a museum or a public landmark and Google sends you information about it. Now imagine a “Robot Goggles” application, Kuffner suggested; a robot would send images of what it is seeing to the cloud, receiving in return detailed information about the environment and objects in it.
Using the cloud, a robot could improve capabilities such as speech recognition, language translation, path planning, and 3D mapping.
The idea of connecting a robot to an external computer is not new. Back in the 1990s, Masayuki Inaba at the University of Tokyo explored the concept of a “remote brain,” as he called it, physically separating sensors and motors from high-level “reasoning” software.
Now cloud robotics seeks to push that idea to the next level, exploiting the cheap computing power and ubiquitous Net connectivity available today.
As a side project, he's now exploring a variety of cloud robotics ideas at Google, including "using small mobile devices as Net-enabled brains for robots,” he told me. "There is an active group of researchers here at Google who are interested in cloud robotics," he says.
But cloud robotics is not limited to smartphone robots. It could apply to any kind of robot, large or small, humanoid or not. Eventually, some of these robots could become more standardized, or de facto standards, and sharing applications would be easier. Then, Kuffner suggested, something even more interesting could emerge: an app store for robots.
The app paradigm is one of the crucial factors behind the success of Apple’s iPhone and Google’s Android. Applications that are easy to develop, install, and use are transforming personal computing. What could they do for robotics?
It’s too early to say. But at the Nashville gathering, attendees received Kuffner’s idea with enthusiasm.
“The next generation of robots needs to understand not only the environment they are in but also what objects exist and how to operate them,” says Kazuhito Yokoi, head of the Humanoid Research Group at Japan's National Institute of Advanced Industrial Science and Technology (AIST). “Cloud robotics could make that possible by expanding a robot’s knowledge beyond its physical body.”
“Coupling robotics and distributed computing could bring about big changes in robot autonomy,” said Jean-Paul Laumond, director of research at France’s Laboratory of Analysis and Architecture of Systems, in Toulouse. He says that it’s not surprising that a company like Google, which develops core cloud technologies and services, is pushing the idea of cloud robotics.
But Laumond and others note that cloud robotics is no panacea. In particular, controlling a robot’s motion—which relies heavily on sensors and feedback—won’t benefit much from the cloud. “Tasks that involve real time execution require onboard processing,” he says.
Stefan Schaal, a robotics professor at the University of Southern California, says that a robot may solve a complex path planning problem in the cloud, or possibly other optimization problems that do not require strict real-time performance, "but it will have to react to the world, balance on its feet, perceive, and control mostly out of local computation."
And there are other challenges. As any Net user knows, cloud-based applications can get slow, or simply become unavailable. If a robot relies too much on the cloud, a problem could make it "brainless."
Kuffner is optimistic that new advances will make cloud robotics a reality for many robots. He envisions a future when robots will feed data into a "knowledge database," where they'll share their interactions with the world and learn about new objects, places, and behaviors.
Maybe they'll even be able to download a helicopter pilot program?
Below are some other examples of cloud robotics projects:
• Researchers at Singapore's ASORO laboratory have built a cloud computing infrastructure to generate 3-D models of environments, allowing robots to perform simultaneous localization and mapping, or SLAM, much faster than by relying on their onboard computers. The backend system consists of a Hadoop distributed ¿le system that can store data from laser scanners, odometer data, or images/video streams from cameras. The researchers hope that, in addition to SLAM, the cluster could also perform sensor fusion and other computationally intensive algorithms.
• At LAAS, Florent Lamiraux , Jean-Paul Laumond, and colleagues are creating object databases for robots to simplify the planning of manipulation tasks like opening a door. The idea is to develop a software framework where objects come with a "user manual" for the robot to manipulate them. This manual would specify, for example, the position from which the robot should manipulate the object. The approach tries to break down the computational complexity of manipulation tasks into simpler, decoupled parts: a simpli¿ed manipulation problem based on the object's "user manual," and a whole-body motion generation by an inverse kinematics solver, which the robot's computer can solve in real time.
• Gostai, a French robotics firm, has built a cloud robotics infrastructure callled GostaiNet, which allows a robot to perform speech recognition, face detection, and other tasks remotely. The small humanoid Nao by Aldebaran Robotics will use GostaiNet to improve its interactions with children as part of research project at a hospital in Italy. And Gostai's Jazz telepresence robot uses the cloud for video recording and voice synthesis.
• At present the iCub humanoid project doesn't rely on "cloud robotics," but Giulio Sandini, a robotics professor at the Italian Institute of Technology and one of the project's leaders, says it's "a precursor of the idea." The iCub, an open child-sized humanoid platform, works as a "container of behaviors," Sandini says. "Today we share simple behaviors, but in the same way we could develop more complex ones like a pizza making behavior, and our French collaborators could develop a crepes making behavior." In principle, you'd just upload a "behavior app" to the robot and it would cook you pizzas or crepes.
[If you know of other cloud robotics projects, let me know.]
Kidd claims that, yes, Autom can help people lose weight. The robot is more effective than weight-loss websites and smartphone apps, he says, because people develop a bond with the robot and stick with it longer.
I think they are onto something here, but I see some limitations in the current robot. First, the speech synthesis is very robotic. Second, the robot has no voice recognition at all. it would be nice if the robot could speak more naturally and if at least basic interactions -- like answering "yes" or "no" -- could happen via voice. The good thing is the company might be able to improve these features in the future with software updates.
Another question is whether consumers want a robotic weight-loss coach in the first place, and how much they're willing to shell out.
Intuitive Automata plans to start selling Autom on its website later this year for around US $500 or $600. But in the video Kidd mentions something interesting: They plan to sell the robot also via health insurance companies and employers, which would give -- or subsidize -- the robots to customers and employees.
Would you take Autom home?
Photo and video: Josh Romero & Joe Calamia/IEEE Spectrum
In our best robots of CES roundup last week, it appears that we left out an interesting offering: the Windoro window-cleaning robot from South Korea.
That's right. This robot wants to do for your windows what Roomba and Scooba do for your floors. It's quite a sight to see this gizmo magically crawling on glass.
But there's no magic, of course. There's magnetism. The robot consists of two modules that go on opposite sides of the window and hold each other using permanent magnets.
Watch how it works:
The mighty iRobot, with its best-selling Roomba vacuums and innovative Scooba floor-washing bots, dominates the cleaning-robot market. But now Ilshim Global, a small firm from Gyeongsan, South Korean, wants to claim a new part of that market -- the vertical segment, so to speak.
The Windoro robot can clean windows 6 to 25 millimeters thick (0.2 to 1 inch). And no, you won't see it hanging on skyscrapers -- its creators say it's designed for cleaning windows at homes and stores.
One of the robot's two modules works as the navigation unit. It uses accelerometers to navigate and bump sensors to detect obstacles and window frames. The other module is the cleaning unit, which has four spinning microfiber pads and a reservoir that dispenses detergent.
The robot first moves up and down and left and right to determine the dimensions of the window. It then follows a zigzag pattern to cover the entire surface, moving at an average speed of 8 centimeters per second and returning to the starting point when it's finished.
One battery charge lasts about 2 hours, and the robot can clean a surface of up to 12 square meters (130 square feet).
The Windoro robot will first go on sale in South Korea, followed by Europe, over the next couple of months. It should be available in the United States in April and will retail for about US $400.
UPDATE 1/19: Corrected maximum surface area robot can clean.
Photo and video: Josh Romero & Joe Calamia/IEEE Spectrum
Back in July, we wrote about how UPenn’s GRASP Lab had taught their quadrotors to work together to grasp and move things. The next step, it seems, is teaching the quadrotors to work together to grasp and move things and actually build buildings. The video above shows a team of quadrotors cooperating to construct the framework of a (rather small) building. The building’s structure is held together with magnets, and the quadrotors are able to verify that the alignment is correct by attempting to wiggle the structural components around, which is pretty cool.
It’s fun to speculate about how this technology might grow out of the lab into the real world… To build actual buldings, you’d either need much bigger quadrotors (which is possible), lots of small quadrotors cooperating in big pieces (also possible), or buildings built out of much smaller components (which might be the way to go). The quadrotors probably wouldn’t be able to do all the work, but they have the potential to make construction projects significantly more efficient.
I would love to have a Scooba, iRobot's floor-washing robot, to keep my kitchen and bathroom shining. But for my New York City-sized dwelling (read: tiny cramped apartment), the rotund robot is an overkill -- it could probably clean the entire bathroom floor just by spinning in place.
It appears that iRobot heard the same complaint from many people and decided to shrink the Scooba. The new Scooba 230, unveiled at CES, is about the same height but only half the diameter of the original Scooba 300 series [see photo above]. At 16 centimeters in diameter (6.5 inches) and 9 centimeters tall (3.5 inches), the new Scooba can get into small areas such as that dreaded space around the toilet.
Like the original models, the shrunken Scooba uses a three-stage cleaning approach: first, it deposits water or a cleaning solution on the floor; then it uses scrubbing brushes to lose dirt and grime; finally, a squeegee vacuum removes the dirty water. The Scooba 230 is designed to clean up to 14 square meters (150 square feet) of tile, linoleum, or sealed hardwood floors in a single session, while the larger Scooba units can clean from 23 to 80 square meters (250 to 850 square feet), depending on the model. Watch the video below to see how the Scooba 230 works:
And how did iRobot engineers manage to shrink the robot and still allow it to clean a sizable area?
The trick is the robot uses the same internal volume to store both clean and dirty water. The two are separated by a flexible membrane and never get mixed; as the clean water goes out, the membrane makes more room for the dirty water coming in from the squeegee vacuum [see illustration].
In terms of navigation software, the Scooba uses the same approach as the Roomba to make sure it covers an entire area, following walls, going around obstacles, and driving over the same spot multiple times -- and sensors below the front bumper prevent it from falling down stairs and other drop-offs.
The new Scooba 230 will be available this spring in the United States and will cost $300. We plan to post an in-depth review soon.
With all of the new competition in the consumer robotics field, it’s about time for iRobot to show that they’re still capable of innovating new and exciting things. AVA, their technology demonstrator, definitely fits into the new and exciting category.
AVA is short for ‘Avatar,’ although iRobot was careful not to call it a telepresence robot so as not to restrict perceptions of what it's capable of. AVA is capable of fully autonomous navigation, relying on a Kinect-style depth sensing camera, laser rangefinders, inertial movement sensors, ultrasonic sensors, and (as a last resort) bump sensors. We got a run-down a few days ago at CES, check it out:
All of the sensor data crunching is taken care of by a heavyweight on-board computer, but the brains of the operation is really whatever AVA happens to be wearing for a head, in this case, a tablet PC. This makes it easy to develop applications to control the robot, which is a concept not unlike the iRobot Create: the building a robot part is done for you, leaving you to focus on getting said robot to do cool stuff.
There are also a bunch of interesting ways to interact with AVA. You’ve got the tablet of course, if you want to do things the hard way. A second Kinect camera on the bot can detect people and recognize gestures, and an array of microphones can detect and interpret voice commands. Finally, AVA’s round ‘collar’ piece has touch sensors all the way around, offering an intuitive way to steer AVA around.
While iRobot wouldn’t speculate on what’s coming next for AVA (disappointing), telepresence is an obvious first application. AVA also has a bunch of expansion ports that you can attach stuff to, which obviously makes me think manipulators. Personally, I’m hoping that now that AVA is out in the open, iRobot will keep us updated with some of the new ideas that they’re playing around with.