This is the sort of thing we do at UAT, at least during "tech forums week" such as the one held in November 2002. This particular contest was hosted by instructor Derric Clark. He got the school to buy four Lego Mindstorms RIS sets and four Accessory Packs.

I was wandering around between classes when someone asked me if I was going to sign up for the robot contest... I arrived late, but yes, of course I would!

Overview printout, showing the teams, and which four will compete on each day.

Sounds simple enough, eh? I was in the last team. At some point Eric Heitmuller changed our team name to SURGE2000. I've divided the details up into each of the three days.

Wednesday

(I don't have the instructions printout for Wednesday.)

On Wednesday I just watched the other teams compete. The challenge presented to them was to knock six randomly placed plastic Lego pucks off a tabletop despite four large obstacle tires.

Bot Heads (foreground), Jon Oser (sitting in back).

Tommy Visic.

During the building period, I watched some teams completely restart their designs multiple times.

Bot Heads, The Empire, Al-Paca, Command Override. The pucks are on the right.

The four Wednesday robots were: a "lawnmower" design, "Tweak-Bot", and two bulldozers.

Al-Paca.

Al-Paca.

Team Al-Paca's robotic bulldozer pushed everything—obstacle tires included—off the table.

Command Override.

Command Override with guard rails.

Command Override's bulldozer got stuck on the tires in the first round, but guard rails were added for the second round, making it an even match for Al-Paca's robot.

Command Override, 2nd round.

Both bulldozers used the light sensor to detect the edge of the table. This allowed them to avoid driving over the edge ...most of the time.

Bot Heads.

Bot Heads, severely overhauled for the 2nd round.

The "lawnmower" design launched the pucks off the table easily when it worked, but it got stuck frequently, and in the final round it put so much weight on the "mower blades" that they barely spun at all. It definitely could have used more building time.

The Empire's robot, AKA "Tweak-Bot".	After an arm broke free.

My favorite was the one I dubbed "Tweak-Bot" because it would twitch spastically when its over-sensitive feeler arms touched an obstacle tire, or the edge of the table, or even one of the pucks... The best part was when all of this random lurching triggered by the arms getting stuck on a tire caused the robot to free itself by dislocating its own arm. The extraneous motorized radar dish only added to the amusement of this spectacle. The irony here is that this robot was based on a model offered in the Mindstorms instruction booklet, yet it proved to be the most broken of the four.

The two bulldozers were the clear winners.

Al-Paca vs. Command Override, head to head, just for fun. It was a tie.

Thursday

Thursday's challenge was only slightly more difficult:

Thursday instructions printout.

Assuming that our robot would at least drive around, I started building a tractor base while thinking about how to retrieve the balls. I thought it'd be fun to give our robot a claw that would close around the base of each ball, presumably triggered by a touch sensor, allowing the robot to drag the ball back across our line...

Ryan Gerhardt, David Shorey, Kenny Fritts.

Loose Parts' "Rake-Bot".

But at about this time I looked over at David & Ryan's team next to us and noted that they were building a giant 3-foot-wide rake. I started to realize that the only way we could hope to retrieve the balls individually, let alone even one of them, using this claw strategy would be to have our vehicle drive around methodically, hoping to bump into a ball, and this would probably take far more time than it would take the giant rake to scoop up all four balls at once, assuming that both robots even worked properly.

Building the "Lasso-Bot".

Scott suggested that we build some kind of hoop or net to drop over the balls, and I wasn't sure if we had enough pieces, but we used all of the flexible tube pieces to form the front of the hoop, and all the girders to form the back, where it was geared to pivot up and down like a big lasso.

SURGE2000 testing the "Lasso-Bot".

Our robot never failed to get all four balls in one approach. Using the light sensor, our program counted the black lines encountered as we snuck forward. Shortly after the second line, our treads would stop as the arm lowered the hoop which was precisely the right size to barely enclose all four balls in their initial positions. Then we'd slowly & steadily drive backwards, dragging the balls across our line. I wanted to add a rear sensor to automatically stop at the edge of the table, but this was not necessary, and Scott didn't want to jinx our working design.

Loose Parts, SURGE2000, The Amazing Flying Pimpsnatch, Mr. Roboto.

Here's the playing field with the four Thursday robots. Beverage not included.

SURGE2000's "Lasso-Bot" was slow but sure.

Loose Parts' "Rake-Bot" with Pete Geiter, Tim Winsky.

David Shorey (fore), Ryan Gerhardt, Kenny Fritts.

Loose Parts' "Rake-Bot" was fast.

In the first round, our robot performed the entire task in 33 seconds. But the Rake-Bot, although haphazardly clumsy in its speed, did the same thing in 3 seconds. It had no moving parts aside from the two wheels, using only its forward momentum to cause the body—rake attached—to tip forward upon braking. It required that somebody remain in position to catch it as it sped backwards off the table with the balls.

The Amazing Flying Pimpsnatch.

Because I was too focused on building, I didn't see the other robots perform much. I think one group gave up because their design proved to be ineffective.

Mr. Roboto.

Brian Clifton, Pete Geiter, Corey Johnson.

Pete Geiter.

I know that Brian & Corey's team built a car with a sort of "basket" on one side with which they were going to scoop up the balls as they drove past, but apparently they had trouble with the programming and/or the weight of the basket causing the car to lean.

Travis has the RIS at home, so even though his team was not officially competing today, he showed up, saw what the challenge was, went home and built a robot with his own set, and came back in time to try it out.

Travis Pabst's non-competing Thursday robot.

Travis Pabst's non-competing Thursday robot with improvements.

He'd built a car which was motored to always turn to the side, with a wheel mounted to roll against the outside edge of the table in order to guide its turn so that it would always drive around the edge of the table. With a basket-arm mounted on the inner side, it would push the obstacle tires off the table and scoop up all four balls in one pass. Apparently it would always lose one ball for some reason, but I was still impressed, as this was totally unnecessary.

After our first round, I decided to re-gear the driving portion of our robot so that it would have less power but more speed. Although it was much faster afterward, during the subsequent test run I was surprised that it did not stop after crossing the second line, and I wasn't able to jump across the table in time to keep it from driving off the edge. I was certain we'd be disqualified for letting our RCX land on the floor, but Derric decided this didn't count because it was only a practice run. I couldn't think of why it didn't stop until I found that the program was written to allow the robot to keep moving forward for 1 second after crossing the black line. In its slow form, it could move maybe 1 inch in 1 second, and this would put it in a better position, but in its fast form, it would travel 1 foot in 1 second. So after repairs and programming adjustments, our competition time improved to 6 seconds.

Friday

So Thursday was pretty easy. I wondered what was in store for the final day...

Final instructions printout.

Ouch, this is just plain mean. Well, we grudgingly set out to build ourselves a bridge, which turned out to be what everyone else did.

This is pretty much what my entire childhood looked like.

I was thinking our robot would be a car which carried the bridge up to the edge of the table and lowered it into a position where it would be released so the car could drive across. But I started to wonder how we were going to get our robot lined up to drive back across again. I suppose we could have used the light sensor or something, but shortly we noted that Eric's bridge was rather flimsy and I tried to consider something different.

Building the "Crane-Bot".

I asked Derric if it would be legal for us to have a second car which was released by the first one, and he agreed that this would be acceptable, as long as whatever crossed the far line could be considered a "robot".

SURGE2000's "Crane-Bot".

So our new plan became to build a crane with a simple arm which would reach across the gap and release a very small "excursion" car which would be powered using the extra-long wires. Then, in order to get back across, the excursion car would be reeled in using the string & bobbin, with the whole thing acting like a fishing pole. This seemed to address all the steps of the challenge.

Eric Heitmuller, Brad Macomber, Scott Richmond, Derric Clark.

Our string got tangled with almost every run and had to be re-spooled.

Testing the "Crane-Bot".

We had to lower our motor power so that the crane wouldn't rise too quickly and cause the whole thing to lurch backwards.

Brad Macomber.

Our crane had its pivot point in the far front, with as much weight as possible behind that point so that it wouldn't tip forward when lowering the arm. A small tower with pulleys guides the string to a high, forward position to make hoisting the arm easier.

Team Loose Parts wisely fitted their excursion car with an antenna to help hold their wire up out of the way.

Loose Parts

David Shorey with a glint of the "Lego Madness" in his eye.

Command Override, with accessory paper.

Jon Oser.

During testing, Jon is ready to catch their excursion car in case it falls off the bridge.

Al-Paca (fore), SURGE2000, Loose Parts, Command Override.

The four finalist teams met once again on Friday, each equipped with some sort of gap-spanning device.

Ryan Gerhardt, Justin Agrell, Brad Sweet.

Brad Sweet.

Team Al-Paca accepted a lower score by intentionally abandoning their powerless car.

Kenny Fritts, Ryan Gerhardt.

Ryan Gerhardt ready to catch a falling robot.

I didn't see what went wrong with David & Ryan's team, but all the teams were overwhelmed with annoying problems.

Jon Oser, Brad Macomber, Eric Heitmuller, Justin Agrell.

Brad Macomber, Jon Oser, Travis Pabst, Eric Heitmuller, Andy Cobb, David Shorey, Justin Agrell, Brad Sweet, Scott Richmond.

This is what it's supposed to look like, though I think this time the excursion car didn't make it back.

Here's a small AVI movie of one of our competition rounds where the car seems to have failed to come free of the arm.

Jon Oser, Justin Agrell, Andy Cobb, Travis Pabst, David Shorey.

Jon Oser ready to catch a falling robot.

Travis' team had the largest, strongest bridge, which they slammed down to cross the gap. Often their speedy car would even remain lined up to get back across, but usually something would get stuck or the car would come back misaligned so that it slipped off the bridge.

I wasn't sure if we were the official winner, because our robot was the only one that could get 100% of the points, but it only worked perfectly maybe 25% of the time, frequently getting tangled in string and wires, and although it would work well several times in a row on the testing table, it never worked perfectly on the competition table even though Derric gave us a couple extra tries.

If only there were more time for adjustments.

Afterward, the school took us out to Bennigan's, and I got a ride there in a certain DeLorean. My team was later awarded gift certificates.

Incidentally, I have come to realize that the UAT students who signed up for this competition are a good sampling of those who I believe show the greatest potential for invention and accomplishment through their participation and willingness to solve problems. Out of all our students, these are the guys who are always found around campus working on their own projects and making things happen. Consider this a hint to any employers out there.