E&T reports from the world’s smallest sports event: a football tournament for micro and nano-robots.
We most readily associate the idea of robots competing at sport with machines built to human-size, or at least human-scale. Honda’s Asimo, Sony’s Aibo and the ‘house’ hatchet-lady Matilda from TV’s ‘Robot Wars’ spring to mind.
The capabilities of these larger machines have advanced at a tremendous rate, keeping them in the public eye.
For example, even though Sony closed down Aibo’s technical programme in 2006, enthusiasts have retained access to its developer kits (in fact, the kits will be supported until 2013). This independent community has since taken the e-dog to the point where kicking a football has become a relatively trivial task – unlike most of the English national side, Aibo can now pass the ball with reasonable accuracy.
So, the ambitious goals set for these more life-sized robots and enshrined famously in the mission statement of the annual RoboCup robotic soccer tournament are looking increasingly achievable.
Let us ask another question. Can we extend the metaphor of sport to the research development and implementation of robotics at scales well below the human or macro, down into the realms of the micro and the nano?
Two teams of researchers from well-regarded US institutions gave up the Memorial Day weekend last May to not merely explore but answer that specific question. The teams came from Carnegie Mellon University in Pittsburgh, Pennsylvania and the US Naval Academy (USNA) in Annapolis, Maryland. The event was the 2008 RoboCup US Open at the Carnegie Science Center, Pittsburgh. The result was arguably the smallest sports event on Earth, but in a good way.
“What we’ve done is get agreement to set up a Nanogram league within the existing RoboCup,” explains Craig McGray, a research associate in the Electronics and Electrical Engineering Laboratory of the US National Institute of Standards and Technology (NIST).
“For 15 years or so, the RoboCup has promoted the football metaphor for research and has been a venue for competition at the macro level. What we’re bringing along now is microrobot and nanorobot technology, things that are coming out of all the exciting stuff happening in MEMS [micro-electromechanical systems] and beginning to mature.”
With NIST as the Pittsburgh event’s sponsor, the teams competed in three football trials on pitches that measured a mere 2.5x1.5mm, with goals 0.9mm wide by 0.5mm deep. Some 16 pitches could be deposited on a single microchip. Electrodes were dispersed throughout the pitches acting as a source of both power and communication for the nanobots.
“We needed to put the devices through tasks to measure three things: speed, control and manoeuvrability,” says McGray.
Speed was addressed by the Two-millimetre Dash, a simple sprint from behind the goal-line at one side of the pitch to behind that at the other. Control was then tackled with a slightly more complicated Obstacle Course. Here, nanobots had to negotiate past a series of ‘defenders’ in the form of polymer posts attached to the chip’s surface. Finally, manoeuvrability was probed in the Ball Handling Drill.
The most complicated of all the tests, this required the nanobot to dribble a ‘ball’ (in reality, a silicon dioxide disc, 50µm in diameter) from the home to the away goal, again through a field of polymer obstacles. Whichever team scored the most goals in three minutes won this particular test.
As these descriptions suggest, the Nanogram league is today more at the training-drill stage than that of fully-fledged competition. Indeed, direct on-the-pitch competition was impossible in Pittsburgh: one of the designs moved in a solution, the other did not.
“However, we’ve already come a long way. This only started back in July 2007 and then you were looking at these robots and sometimes they wouldn’t move at all. Now, just a year later, you’re seeing directed movement,” says McGray.
“All the same, we are being straight with people. We are not working towards the China RoboCup this year, but 2009 in Graz in Austria.”
What really matters, though, is that the Nanogram challenge and the football context help to encourage some interesting research. There’s evidence that they have. The metaphor really does reach down that far. Both devices were approximately 300µm in size, although as noted earlier, they then applied differing technologies.
Carnegie Mellon’s entry was machined from a magnetized rare earth material, neodymium iron boride. The device was manipulated by the pitch’s embedded electromagnets. Direction and speed were determined by the direction and magnitude of the magnetic wave pulse that reached the nanobot.
The USNA produced its device from silicon and chromium. It moved through a solution in response to an electrical waveform sent across the field of play. Turns were negotiated by causing a stylus, in the form of a wing-like tip on the side of the structure to the main surface of the microchip.
“What you get to look at are foundation technologies for some application areas with high potential,” notes McGray. “We’re looking at devices that might ultimately be small enough to suspend in a solution so they pass through a needle and get to work in the bloodstream, for example.”
This hints another way in which the Nanogram league could diverge in fascinating ways from the mainstream RoboCup community. While the existing competition is married to human and other animal forms, and to applications that are interactive at the macro scale, the Nanogram devices could go wherever they like.
“Bill Gates recently said that robotics was going to be one of the big technologies of the next decade or so, and I don’t think he was just thinking about the traditional view,” says McGray. “I think he had the micro and nano scales in mind, when he said that.
“The existing RoboCup has done a lot to raise people’s awareness at the macro level. The great thing about it and about comments like those of Gates is that they now give us a chance to do the same at levels beyond that.”
With Germany’s reputation for ruthless efficiency on the football field, it is unsurprising that their scientists should design a robot goalkeeper.
Goalias is the product of six months work by a team from the University of Stuttgart’s Institute for Automation and Software led by Prof Peter Goehner (CORR).
The goalkeeper itself is no more than a cardboard cutout which moves along a track between the goal posts.
It is when a player strikes the ball that the engineering behind Goalias comes into play. Three cameras on the crossbar send images to a computer which analyses them and then directs the keeper.
Goalias was developed as a way of getting young people interested in engineering.
Prof Goehner told E&T: “Over nine days we had 280,000 people visit the exhibition and many of them were young people. We did not think Goalias would be so popular.”
He has also been very good at his job. Out of 11,000 shots taken from 11m at a full-size goal, the robot has saved 91 per cent. Even when top footballers took up the challenge, he managed to save 50 per cent.
Prof Goehner said that getting Goalias to work required three key components. Recognition software to analyse the pictures, a microprocessor to calculate where the ball was going and the actuator to move the keeper at up to 60kmph.
Getting Goalias to the right spot in time means assessing and acting on the data in under 400 milliseconds.
The success of Goalias has led to interest from companies and sports clubs. Prof Goehner believes that it could be adapted for different sports as well as inspire other uses for its speedy automotive system.
He also believes it can get even better between the posts.
Further information:
www.ias.uni-stuttgart.de/goalias/start.html
The relationship between football and robotics has not typically been a friendly one.
Englishmen of a certain age still shudder at the mention of ‘Revie’s Robots’. They were the successful-but-widely-loathed Leeds United team of the 1970s, one drilled in football and thuggery with parade-ground precision by its manager, the eponymous Don Revie. More recently, England and Portsmouth striker Peter Crouch picked up the ‘robotic’ tag at the 2006 World Cup for his celebratory ‘robot dance’.
Researchers in the scientific world have addressed the relationship from the opposite point of view. For them, linking robotics and football has obvious benefits.
Football requires a combination of speed, control and dexterity – all being executed with grace and a considerable degree of anthropomorphism – that is intrinsic to many of the larger scale applications robotics researchers are targeting in the medium- to long-term. Such applications are also likely to depend on various agents sharing information and acting like a team under adverse and confrontational circumstances. Examples range from robotic nurses to military probes.
The success of TV shows like ‘Robot Wars’ and the fact that the number of children playing football is growing suggest that intelligent machines playing the world’s most popular sport could capture the imaginations of tomorrow’s potential engineers. This is partly about making technology cool.
The event that best reflects this ‘positive’ view, RoboCup, was proposed in 1994 by Hiroaki Kitano, a member of the AI development team for Sony’s Aibo robot. Since 1997, it has become an annual and global robotic soccer tournament, comfortably beating the human World Cup in frequency.
As a result, RoboCup (officially but seldom referred to as the Robot World Cup Soccer Games) has a well-established infrastructure into which new scales of robotics, such as those featured in the Nanogram league, can be introduced. Tucker Balch, director of the Institute for Personal Robots in Education at the Georgia Institute of Technology, points to two further benefits of RoboCup’s increasing maturity.
“First, researchers understand the aim of the technology, so it’s easy to get them interested. Second, people become emotionally involved because of the competition,” he says.
Balch underlines his own enthusiasm for RoboCup when asked if he follows a human club. “Sorry, no. Compared to RoboCup soccer, human soccer is boring,” he answers, with a wink.
Like this year’s Olympics, RoboCup 2008 takes place in China, albeit some way from Beijing. Competition began on 14 July at Suzhou in Jiangsu Province, a designated high technology zone in the east of the country.
RoboCup’s quest for a combination of physicality and grace is reflected in Kitano’s original and ambitious mission statement.
By the mid-21st century, a team of fully-autonomous humanoid robot football players shall win a football game, complying with the official rules of FIFA, against the current World Cup holders.
RoboCup may represent the best chance a British team has of claiming world footy domination in our lifetimes (although we would have to beat current robochamps Germany, Iran, Japan and the US first), the growing viability of its end-target reminds us how far robotics has come since being primarily associated by the public with the assembly line.
Bookmark with:
del.icio.us
Digg
Add to Discover
Comments
All comments
You need to be registered with the IET to leave a comment. Please log in or register as a new user.