The RoboGames are returning to San Francisco in April. Hey, Robots, you're on the same team!
RoboGames, now in its 7th year, is the world’s largest robotics competition. For this year’s contest they secured the 5,000 person San Mateo Expo Center, just outside of San Francisco. From April 23-25 this will be the hallowed grounds where thousands of contestants, controlling more than 500 robots, compete in 70+ events. Many of these events will be robot-on-robot battles such as biped kung-fu and robot sumo wrestling. There will also be an exciting new event this year called “mechwars.” Contestants will battle using walking robots, but their view will be restricted to a bot-mounted pov camera. They will then compete in a scale-model city with flamethrowers and CO2 powered rifles to take out their opponent. How hardcore is that! I was able to talk with founder David Calkins and get the scoop on how that competition and others makes the RoboGames better every year.
Most events will take place in the shadow of the biggest draw at RoboGames- robotic combat. In these games, robots go after each other in a lexan-enclosed arena with everything from buzz saws to pneumatic hammers.
But RoboGames is about more than just honing the combat capabilities of our future overlords, the competition includes robot marathon and sprinting competitions, hockey, soccer, basketball, and junior events. In talking with RoboGames founder David Calkins, he said that events will continue to be added going forward based on what the robot designers are interested in. In the past, this meant adding the many artistic categories that are now part of the games, “we found we had a lot of requests from robot artists to compete.”
On a side note, I’ve always wondered why robot sprinting isn’t a bigger deal. The 100m dash is probably the marquee event at the human Olympics, but sprinting is just another event at the RoboGames (actually, it’s three events, for different sized robot sprinters). But I’m not blaming the organizers, it just seems like nobody takes the “world’s fastest robot” title seriously (of course at SingularityHub, we do). If you Google “world’s fastest robot” you get videos of quick working sorting bots, not Usian Bolt wannabees. Organizers of the yearly RoboCup, where humanoid robots play soccer, have stated their goal to field a robot team that can beat the human winners of the World Cup by 2050. That’s a pretty distant goal (mind the pun), but it’s distant for a good reason- playing a good match of soccer would required just about every aspect of a humanoid to be perfected, and then some. Beating the world’s faster runner to 100 meters seems more like it’s within the realm of possibility, at least something that a researcher could both start and finish. Plus, it just looks cool.
Toyota’s humanoid running robot, the world’s fastest:
Last year, teams from all over the world came to RoboGames and this year will be no different. Newcomers from Pakistan and Egypt will join traditional powerhouses like Japan and Taiwan, as they go for the gold. The games are a huge deal for many of the foreign teams, some of whom publish research papers based on their experiences at the games, and a few, like winners from Mexico and Indonesia last year, get an invite to meet their country’s president.
This international flavor goes a long way towards accomplishing one of the main goals of RoboGames- strengthening the bonds between different groups of robot engineers. Many of the events at RoboGames involve only a small subset of robotics. For instance, combat bots are very mechanical engineering heavy, whereas the automated events are often very computer engineering heavy. Other events have strong electrical engineering, or artistic components. David thinks the games help cross-pollinate the various fields with information from other researchers.
But cultural exchanges, spirited competition, and gold medals aren’t the only thing that the RoboGames share with the Olympics. When I asked David how the competition changes from year to year, he told me “the games are pretty much the same, but the contestants just keep getting better.”
The 7th annual RoboGames will be held from April 23rd through the 25th. For more information, head to http://robogames.net/
The San Mateo Expo Center is also the location for the Maker Faire, coming May 22nd. More on that, later.
[photo credit: Kaist]
[video credit: Blogphilo New Media, Toyota via Smart Machines]
Comments
You guys make some good points. However (you knew that was coming)…
The funding situation is actually more rosy than it seems. It's true that not much money will be invested in the particular application of robotic sports. But the general application of robotics as a substitute for human labor has attracted a lot of attention and dollars. The Japanese and South Korean governments need to address the impending shortage of younger workers, brought on by lower birth rates and increasing longevity. In America, the solution would be to admit young immigrants. In Japan, however, they are turning to robots.
We've been reading for a while that Japanese companies are scrambling to develop machines to take care of the elderly. Cooking, cleaning, bathing, and companionship will be some of the services being targeted. In order to perform these duties, a robot will have to navigate an unstructured environment, manipulate objects, and use some basic forms of common sense.
In the US there are also many strides being made. Robotic cars can already drive themselves in some fairly tough environments (they'll challenge Pike's Peak this year). And Willow Garage, among others, is producing robots able to open doors and plug themselves into wall outlets. The military already has machines with some limited autonomy. Watching a video of Big Dog running over snow and uneven terrain makes me feel that capable two-legged robots, though more challenging to build, are not as far off as people may think.
Sure, the current robotic vacuum cleaners and lawnmowers are a far cry from the autonomous servants of science fiction. But competition in those industries is starting to heat up. Once robots become capable of holding a broom and clearing the table (only a few years hence), the demand for them will soar. That will spur tremendous development efforts, much like the smart phone industries are experiencing today.
My point is that the science and development of robots will advance quickly because of their usefulness in many diverse fields. In fact, most of the cost of goods and services today is determined by their labor content (which is very evident from the practice of outsourcing). This provides a huge incentive to replace human labor with capable machines. And once a robot can assemble a product more cheaply than can a human (also a few years hence), we will permanently shift that job to the machines, which will only become even cheaper and more capable the following year.
I think this decade will see monumental increases in robotic capabilities. And as an offshoot of these benefits, robots will eventually possess the physical prowess exceeding that of unenhanced human athletes.
I'm afraid I have to disagree with you on this one Daryl. I know this sounds weird, but I honestly think a computer will pass the Turing test before a team of robots wins the World Cup.
The biggest reason: research funds. Untold billions of dollars goes towards creating faster processors, very little, comparatively, goes into making human-style robots. This accounts for Klas' correct impression that not much has happened in the robocup over the last few years. Or, at the very least, certainly nothing like the exponential growth we see in computing.
This research disparity is magnified by the challenges involved. Even if you could get a robot to run in a straight line faster than a human (a very tall order by itself), you would have to adapt that robot to be able to make fast cuts on a slippery surface, while also having incredible dexterity with their feet (imagine the difficulty getting servo motors to act like this guy: http://www.youtube.com/watch?v=KZnUr8lcqjo). On top of that, you would need to add a very powerful source of energy to fuel those (relatively) inefficient servo motors, plus a lot of thirsty computer chips, for 45+ mins.
I'd love to be proven wrong on this, but I think humans will have this one for some time (and maybe forever, given the potential for augmentation).
I hope you're right, but I'm not so sure. I remember watching robocup videos from both -06 and -09, and I got the impression that the skills of the robots haven't improved very much in these three years. Perhaps I shouldn't think linearly though, but exponentially.
I've always thought it very conservative to target 2050 for the goal of a robotic team beating the human World Cup champions. Sure there are huge technical challenges to be overcome before robots can play soccer at anything approaching a human level. But keep in mind the exponential progress being made in all areas of this field — mobility, balance, sensors, navigation, computer vision, swarm intelligence etc.
I think getting machines to play human-level soccer is a much easier goal than passing the Turing test. My prediction is that a robotic team will be able to beat the 2022 World Cup championship human team. This assumes, of course, that the human players will not be allowed any non-biological augmentation. Any other opinions on this?