6.270 Autonomous Robot Design Competition Main 6.270 Page
OUR BIG BABY
Michelle Nadermann
Janice Kim
Team 45

So, our robot didn't really make it to competition. The name fits it pretty well. By the end, all it could really do is crawl and wander aimlessly. It made some pretty funky noises too.


[ description | strategy | code | pictures | mistakes ]

Description

Our robot stood about 8 inches tall, and the solid part of our base took up about another 7 by 10 inches. We used a total of 4 tires. On the front side of the robot, we attached the most flexible of our lego parts. These acted as guides, so that we could move our robot behind an "egg," push it forward, and not have to worry about the ball rolling off to the side as we pushed to towards the trough. Including these flexible guides, our robot's entire base took up about 1 square foot. Our battery pack was mounted on top of our gear boxes. On top of our battery pack, we put our trusty handyboard. If we could have redone our design, we might have distributed the weight a bit better. Stacking our heaviest components in this way caused a lot of weight to be put on the rear tires.

Motors: We had two motors attached to gear boxes that powered the two rear tires.

Servos: Our Big Baby had power steering! We attached one servo to each of our two front tires.

Sensors
We used a total of 6 sensors.

  • 4 Digital Sensors - The base of our robot was roughly in the shape of a rectangle. We embedded 4 sensors on each side. The front sensor was about 1/4 of an inch above the ground. The other three were a little less than 2 inches from the ground.
  • 2 Analog Optical Sensors - These two were attached near the bottom, rear corners of the robot.

Strategy

Our strategy was a simple one. First, we would orient ourselves towards the ball nearest to the trough. To do this, we used the optical sensors to figure out which side of the board we were on. By rotating in place, we could figure out where we were in relation to the trough.

Once we oriented ourselves, we moved straight forward to push a lone ball into the trough. The front sensor told us when we hit the 1/4 inch barrier in front of the trough. After doing this, we simply moved back and forth, parallel to the trough in a blind attempt to be a bother to our opponent robot.

Originally, we planned to do more than this. This was just the first step in a long series of steps to push as many balls as we could into the trough. However, due to time constraint, this is as far as we got.

Code

This is where our team failed. Due to some unforseeable problems not relating to the class, we didn't have as much time as we would have liked to develop and test our code. Although you're probably better off not seeing it, you're welcome to take a look: chicken.c. It's written in ic (interactive c). It's very uncomplicated and not very robust.

Pictures

Click on a thumbnail to see a bigger version.
pic1 pic2 pic3 pic4 pic5

Many, Many Mistakes

Doh! It was pretty embarassing how many mistakes we made. Here are a few:
  • We blew our beacon... oh yeah. This came from not reading the little "+" and "-" signs on our beacon and expansion board. This was a big oppsie on our part. A big thanks goes to Jan for taking the time to fix it for us.
  • We originally wired our batteries in parallel instead of in series. Before we did any serious damage with these, a friendly TA informed us of our mistake, and we quickly fixed it... phew..
  • We short-circuited our battery a few times and nearly set off the random hall fire alarm. We have an extremely crispy set of wires that give testimony to how dangerous these batteries really are. We eventually learned our lesson. Batteries with lots of current = DANGEROUS.

P.S.

We had to find some way to get our pictures from the digital camera to the network. ebay turns out to be great for that. Hopefully no one will be interested, as we set our minimum price pretty high, because despite its character flaws, ya can't help but get attached. (though if they're clever enough to buy it for its $1000+ worth of parts... doh!) Check it out: here

email us: gorpwarp@mit.edu, jjkim@mit.edu