- About Us
- Nano-Social Network
- Nano Consulting
- My Account
From nanobots to walking humanoids, robotics technology is rapidly becoming more diverse and increasingly sophisticated, as demonstrated at the 2007 RoboCup.
July 17th, 2007
I recently spent several days at Georgia Tech for an annual event called RoboCup. This is an international research and educational initiative whose goal is to develop, by 2050, a team of fully autonomous humanoid robots that can play against the human world soccer champions—and win. The exhibition is currently separated into five leagues which address various aspects of robotics development. Teams play against each other, culminating in a winner for each league. The leagues range from simulation (basically computer-generation) and wheeled (think R2D2 from Star Wars) to four-legged (we're talking Sony AIBO here) and humanoid. The latter are robots that balance on two legs, just like a person.
You might have seen some news stories recently about the event on CNN and by the Associated Press; they seemed especially intrigued by a new division that was introduced this year—which was called the Nanogram league. Here, micromachined robots (basically tiny MEMS devices) were meant to demonstrate rudimentary soccer skills, like maneuvering around an obstacle and "kicking" a ball into a goal. The field of play was just 2.5 mm x 2.5 mm, with the goals measuring 900 micrometers wide by 500 micrometers deep. Needless to say, you need a microscope to see stuff this tiny, so this was projected onto a large, wall-sized screen to watch all the action.
The soccer ball, which was also micromachined, was about 100 micrometers wide (about the diameter of a human hair) and just a few micrometers high. The microbots themselves were larger than I expected (typically a few hundred microns across) and I was surprised to see that two of the four teams chose to use electromagnetics as a power mechanism; the most prevalent power choice for MEMS devices by far is electrostatic. Both approaches have their pros and cons in this kind of situation, so it was interesting to talk with the teams and find out why they chose the approach they did. Sponsored by NIST, the contenders were from Carnegie Mellon University, Simon Fraser University, the Swiss Federal Institute of Technology and the U.S. Naval Academy.
In theory, the concept of the nanogram league was seriously cool. But, the intended drills didn't quite work as planned; clearly there's a lot of work to be done before this league is really up and running (no pun intended). The biggest challenge was simply to get the bots to move around the field (there were considerable stiction problems the day I watched), much less navigate them with any semblance of control in a specific direction (despite the use of acupuncture needles to nudge them along and point them in the right direction). But that's ok. It was a great demonstration of not only what's possible, but the difficulty of working at this scale. I can't wait to see how participating teams solve the challenges for next year—it's only going to get better.
I spent the majority of my time at RoboCup watching two humanoid leagues: kid-size and teens. The teen league is comprised of two-legged robots that are about 4 feet tall. What struck me was how incredibly slowly (and methodically) they moved. But, they were able to both successfully kick (and defend) goals, although defense was more the result of a misdirected kick than any real defensive play. Despite the slo-mo action, the bots apparently work hard and play hard. It was a bit disconcerting to see one of the robots lying on its stomach in between rounds as team members fanned it with paper—presumably to cool off overheated circuits—but it was evident that a robotic version of heat stroke was in play here.
The kid-size league (two-legged robots about 2 feet tall) was, in a word, hilarious! Once the ball was placed on the field, a few just stood there and simply looked at the ball, then up and around the field, and then back at the ball, as if wondering—what the heck do I do with this? The visual recognition was clearly there, but it was obvious that eye-foot coordination (or software programming in this case), still needs a bit of fine-tuning.
Although it looked like these robots were head-butting and karate-kicking as part of purposeful defensive moves, it was simply a matter of loss of balance—despite the use of gyros and accelerometers to stay upright. Most of the robots quickly marched along in tiny little steps to position themselves; at which point they would carefully lift a leg up and back to the give ball a whack; but those extra seconds often meant they toppled over, frequently taking their opponents with them. If that wasn't funny enough, the robot "goalies" even had the sliding dive perfected to defend the goal against a soccer ball—unfortunately, the timing was almost always off and so the dramatic move often took place after a point was scored.
Everyone rooted for them anyway. I've never seen parents cheer as enthusiastically at their children's soccer games as the audience did for these diminutive robots. It was a blast.
Beyond the sheer fun of this event, it's clear that the level of sophistication found in robotics continues to move forward by leaps and bounds, with MEMS gyro sensors and accelerometers playing an important role in balance. However, a lot of work certainly remains to reach the goal of a world champion soccer team being beaten by their robotic counterparts. Just wait until the electronics are in place; then it'll be cool to see nanomaterials come into play to give these guys human-looking qualities. Good thing we have nearly 4 decades to work on all of this; although, from what I've seen, I suspect it may not take that long.
This article is a transcript of the Bourne Report Podcast #53.
Want to know more? Listen to the weekly podcast:
For more information, please visit: http://www.bourneresearch.com
© 2007 Bourne Research LLC. All rights reserved.