IAP 2000 MIT 6.270 "Bots in Blue"

Team 38

"Piece of Marklar"

I-hsiang "Very Excitable" Shu, Beau "LegoMan" Tateyama, and Laura "Coding Beastette" Currea

This is our "Piece or Marklar"

The Marklar

If you haven't figured it out already, our marklar's marklar came from the television marklar "South Marklar".  The Marklars are featured in the marklar of our marklar, as you can see.

Because you are probably not from Marklar or don't understand our Marklar, we will translate for you...

If you haven't figured it out already, our robot's name came from the television show "South Park". The Marklars are featured in the background of our webpage as you can see.


in the words of Eric Cartman:

 "The Stratejah"

Our original strategy was the following: put the first hacker into our jail; go along one of the sides of Mass. Ave, somehow shoot one or both of the professors on that side onto our campus; cross over to the other campus, and bulldoze the opponent's student and hacker blocks over to our side (for a net of 2 points), and while returning to our side, bringing the last two professors with us, and maybe putting the other side hacker in our jail. If completed successfully, it could have given us a 22-point advantage. However, due to technical difficulties and lack of time, we were unable to complete a robot capable of doing all these things.

After much deliberation, and many changes of mind, we finally decided on a semi-aggressive strategy. Our robot was built with a block shooter on the bottom, so that P.O.M. could go along the back wall and shoot the first hacker into our jail without looking back. Then, he'd make a 90-degree turn and follow the side wall to Mass. Ave. As he passed through Mass. Ave., our robot would close what we affectionately called the "Poop Shoot", and scoop up the professor, with the intention of taking it back over to the opponent's jail. On a flawless run, it could have scored a net of 14 points. Unfortunately, our robot was never able to complete this strategy successfully. But it did have some pretty cool features, described below.

 "Schweet Features"

Wall following...i.e., The Wings

In order to travel quickly across the board, we decided to use wall-following. Other navigation methods, such as line following and shaft encoding, were found to be ineffective at the speeds we wanted to travel.

To wall follow, we mounted gears horizontally on wings that could be raised and lowered. These wings were sized such that when they were down, the horizontally-mounted gears would roll on the wall, guiding our robot straight against it. This allowed us to move in straight lines against the wall very quickly.

Wings up

Wings down

Following a makeshift wall

The wings were sized such that our car's center would be exactly 9" away from the wall, aligning it with the placement of the professor and hacker blocks. This was necessary for our shooter (see below) to work properly.

The wings were actuated by a servo and wires tied to each wing. It was a last minute revision to our car. Originally, the wings were just rubber band loaded and dropped at the beginning of the contest.

The Block Shooter

In order to quickly move the hacker block into the jail and collect the professor, we built a block shooter mechanism on the bottom of our car. It worked by spinning two horizontally-mounted wheels in opposite directions, sucking blocks right under the car.

The Poop Shoot

To collect the professor block to take into our opponent's jail, we needed an additional gate to stop the professor block from shooting through our car. This was the poop shoot. It was a servo-actuated gate at the back of our car. At the start of the round, it would be open, allowing the hacker block to pass beneath the car into the jail. When we went for the professor it would go down, stopping the professor from shooting through our robot.

Poop shoot up

Poop shoot down

 

The Butt Wiggle

One unexpected problem we had was getting stuck on the wall after making 90 degree turns. To compensate for this, we came up with a unique solution. At first, we would just turn and get stuck on the wall. Then we would pulse the wheel on the side of the car that was stuck to the wall forward while driving the other wheel in reverse until both back bump sensors hit the wall. This resulted in a strange wiggling motion our robot would do every time it backed into wall while turning. It was very effective in aligning our car in a reference direction and fun to watch.

before turn
after turn
"the wiggle" full steam ahead

 

Side Block Collectors

Another unexpected problem was the 4 blocks centered at the cross-hatch in front of Mass Ave. We didn't want to take them into our opponent's side and give them points and we were too wide to push them out of our way. So to work around this we made collection chutes on the side of our car where we could hold the 4 blocks, turn, and leave them on our side of campus.

<insert pic>

Speedy Drive Train

In order to make our car have a good top speed and an excellent acceleration, the drive train was built to support 6 motors. In testings, this gave our car incredible acceleration and a very respectable top speed. The car actually did wheelies when we started it from rest and would hold the wheelie until we slowed down. It was a real speed wheelie unlike the fake, lame artificial wheelie done during the exhibition round during contest night.

We used a gear ratio of 27.

 

 Results... super weak

Round 1

During our practice runs, we were ocassionaly able to shoot the hacker into our jail. It worked, maybe 25% of the time. We had a bug we didn't know how to fix. During round 1, to no big surprise, we didn't score. We just banged ourself into the corner and did a few exhibition circles.

Qualifying Rounds

After much head scratching, we finally figured out how to align ourselves with the wall correctly (see butt wiggle and wings above). During practice rounds we were able to shoot the block 90% of the time, but during our first two attempts to qualify, we had problems (the very excitable Shu had some ishues to work out).

But finally, we got it right and qualified! Super Schweet...

Round 2

Poo happens. Our car had stage fright and didn't perform. It didn't shoot the hacker block far enough back to reach the jail, and got lost on it's way to Mass Ave. We're still not sure what happened...most likely a sensor calibration problem.

 

 Problems Encountered

 

 Final Thoughts

Beau

I can't help but say that I was VERY disappointed with this class. It had so much potential to be cool, but ended up just sucking hard. The organizational problems and faulty hardware (board) were just too much to put up with. This year 6.270 was horrible. I hope the organizers do a better job in the future getting contest together and secure a better board.

On the upside, the contest theme was good and the contest board was satisfactory.

Laura

I don't think I was as disappointed with this course as my teammates. Although I do wish the organizers and TAs had put some more effort into preparation, I do not think the experience was without it's good points. I had fun writing the code that made our robot come to life, and discovering cool ways to correct its flaws. I probably won't ever do anything like this again, and am glad that I had the chance this IAP.

I-hsiang

I must say towards the end of this class I was finally starting to get excited, when we actually had working boards and could test our robots and our ideas to the fullest. It was a lot of fun seeing our little creation journey through the course. Although many times it would get lost, it was all a matter of time before we could have had an awesome robot. But there was no time. I really hope next year, someone gets their act together and actually prepare for this great contest. Sometimes I wonder how much of the class was actually ready before January 3rd. I know it's all volunteer work but still... it's an MIT course.

 

More 6.270 Pictures

Return to 6.270 Homepage