|First installment of
the adventures of the Scarabs RoboCup Rescue team at
RoboCup 2002 in Fukuoka, Japan, as reported by Team
Coordinator Michael Randall.
This account is fairly detailed and a bit technical -- not with the intent of boring you (although it may...), but to convey just how much we had to go through to get our robot on the field.
Finally, please remember that this is a team of high school students, acting as the sole representatives for the United States in the Robot Rescue League, competing against university / government / corporate-funded teams of Ph.D.'s and grad students, in an event designed to help save lives.
Our first real taste of Japan! It took us a little while to figure out the currency... and communication. We were AMAZED at the cleanliness (for example, we saw one white-gloved worker scrubbing a chair, with a supervisor on either side inspecting his work!). The level of service also impressed us. (Ex: someone at every escalator to help with your baggage carts - which could ride on the escalators!).
On to Fukuoka! An hour's flight, and we were there! Stepping out of the terminal looked just like a scene from Blade Runner: lots of neon! Fukuoka's excellent subway system whisked us to within two blocks of our hotel. (The HARD part was lugging the 70 lb. suitcase up four flight of steps to reach street level - we didn't discover the elevator until later...). We found our hotel, checked in, unpacked and COLLAPSED!
That's when the trouble started...
On opening the suitcase with the robot, we discovered a number of the components had been damaged in shipment. Most damage was easily repaired... BUT... one of the drive train gearboxes had been forced open, and one of the gear shafts was missing! Of course we had no spares (not something we anticipated would fail...). Fortunately, Josie noticed that some of our pop-rivets had the exact diameter as the missing shaft. So, after a few minutes' filing, we were back in business!
On to the next task: assembling the slip ring on the cable reel. (For those not familiar with our robot, the plan is to carry a supply of Ethernet cable on the robot, laying it down and reeling it back in as needed. A slip ring is a series of moving contacts that transmits the computer signals from the stationary camera to the rotating reel). In principle, a slip ring is very simple; in practice, we had a tremendous amount of difficulty implementing our design in the cramped space available on the robot. (Josie pricked her fingers at least twice while shortening and installing the springs to tension the slip ring contacts).
Not that that was the only issue: we still had to get the motor running to turn the cable reel. We brought electronic components with us to finish the motor driver, but were concerned that our Radio Shack parts weren't really what we needed. But hey, we figured "we're in Japan - land of the transistor". So Kenichi san (our favorite event staff member) took a small group of Rescue Leaguers (including Ian and myself) to Best Denki ("denki" means electronics in Japanese). More consumer electronics than you could concoct in your wildest electric dreams... but not a chip in sight. (We DID buy an Ethernet hub and extension cable to help deal with ANOTHER issue -- more on that later...).
We hopped on the subway and headed for the REALLY big Best Denki in Tenjin. (12 floors!! Made the other Best Denki look like a TV repair shop...). EVEN SO, they had never heard of the optoisolator chips we wanted! So, we bought a few other needed components, and started the long bus ride back to the Dome. (A three-hour excursion altogether... we were NOT popular with the rest of the team...).
Kevin built the reel driver AND... it didn't work. Not his fault - we were experiencing weird voltage spikes, apparently due to mixing the 5 volts of the Stamp board with the 7.2 volts used to drive the reel motor. A simple solution would have been to use an optoisolator chip to separate the voltages, BUT when we tried the Radio Shack chips we brought with us, they didn't work (wrong type). We already knew the good people at Best Denki had never heard of an optoisolator, so we got on the internet to find manufacturer names and part numbers (reasoning that it might be easier to get what we wanted if we were more specific...). Another trip to the Tenjin Best Denki - no luck! They couldn't cross-reference the chips. So we initiated Plan B, and bought a 5 Volt relay (we could use the 5 V signal from the BASIC Stamp to switch the 7.2 Volt motor on and off)... or so we thought.
The current output of the Stamp wasn't strong enough to turn on the relay, so we used a transistor to boost the signal. IT WORKED! The Stamp actually turned the motor on and off via the relay. Then... it DIDN'T work! (AAAARRRRGGGGHHHH!) Seems we were still getting mysterious voltage spikes, which FRIED the relay. (ANOTHER trip to Best Denki...). We installed capacitors with the new relay - THAT seemed to take care of the problem.
only other task was to modify the BASIC Stamp code to be
compatible with the relay. (Originally, it had been
written to control the reel speed by rapidly turning the
motor on and off - something relays don't like to do...).
Since our programmer Valerie (Josie's mom) was not able
to attend, Ian and I had to tough out the code.
Eventually, we made (what we thought) were two simple
changes. (Valerie, I can hear you laughing from
We still had the slip ring beast to deal with. Josie, Kevin and Lynn were ready to mutiny over the design, and I can't say I blame them.
By the way, while all the business above was going on, KIDS were walking through the setup areas. LOTS of kids! BUSLOADS of kids! All well-behaved and fascinated with what we were doing. Turns out, the general public was allowed to tour both the rescue courses and our work areas with little restriction. (This went on for the entire competition). I'm sure the kids enjoyed it, but it was very distracting -- and unprecedented: previous competitions had the work areas off-limits to the public. A good thing, if for no other reason than public safety (hey, we had hot soldering irons sitting on the table...)
We took the cable reel and reel motor off, and configured the robot as a cable-dragger. Which it did just fine...
At this point, I want to say a little about the funky scoring system they use in this competition. Without going into gory details, the equation used favors course difficulty, multiple robots and accurate mapping, while discouraging course disturbance / damage and operator control (i.e. your score is divided by one plus the number of operators TO THE THIRD POWER. For example, if you have two operators, your score is divided by (1+2)^3, or 27!) Their ultimate goal is NO operators: robots smart enough to go unsupervised into a disaster scene and locate survivors. Maybe someday...
While I'm at it, here's a little on the rescue course itself: three separate mazes: yellow (easiest), orange (intermediate) and red (close to a real-life disaster... or my room). Teams are NOT allowed to see the course prior to competing, on pain of disqualification. Two major (unintended, unannounced) changes in the course were made from last year: ramps and wide-open spaces. These proved to be our undoing (more on that in a moment...).
Back to the competition! The referee gave Ian a countdown, and he was off! And the troubles started. Unlike last year in Seattle (where the robots were placed at the entrance of each maze), Ian had to navigate an open space over 40 feet long to reach the yellow maze. The omnidirectional optical system on our robot is lucky to see four feet before the image gets too small; this is great for navigating tight, cluttered areas, but not open spaces. Long story short, he got lost. After a while, he called Kevin in as a second operator to help him figure out where he was. In time they DID find the ramp to a maze. Thinking it was the yellow maze, they started up... then rolled back. Our little robot was originally designed to run on a flat course with a lighter load - the motors couldn't do it! Ian backed up to get a running start.. and ran into the crowds! (Remember earlier, I mentioned they allowed the public in? There were crowds of people lining the course route, held back by rope barriers). People scrambled to get out of the way, but it didn't help much. The robot simply didn't have enough power for it's weight.
And so, we ended Trial 1 with a score of "0" points.
you'll have to wait for the next installment to find out
how it all turned out...
Send your questions about RoboCup to: Imagiverse - Ask The Expert
5 January 2003
© 2002-2003 - www.imagiverse.org