2.007 Journal

[ZZII 527]

Spring semester of sophomore year in MIT Mechanical Engineering has come to mean one thing: 2.007. This is Design and Manufacturing I, a robotics competition disguised as a twelve-unit class which every mechanical engineering major must take. For anyone not familiar with the history behind it, it used to be known as 2.70 and used to be taught by none other than Professor Woodie Flowers. It should be no surprise, then, that 2.007 resembles FIRST, or, since 2.70 came first, FIRST resembles 2.007. In either case, they are very similar in concept: a different game field every year on which robots compete against each other. But they are also very different in scale and complexity. A quick comparison:

COMPARISON: 2.007 / FIRST
--------------------------------
Field Size: 8'x8' / 54'x36'
Format: 1v1 / 3v3
Match Duration: 45 sec. / 2 min. 10 sec.
Robot Size Limit: 16"x16"x26" / 28"x38"x60"
Robot Weight Limit (minus battery): 10 lbs. / 120 lbs.
Control: direct wiring (PIC optional) / PIC-based
Build Time: 13 weeks / 6 weeks
--------------------------------

So, in many ways FIRST has grown larger and more complex than it's predecesor. I have always been amazed that high school students in FIRST get the opportunity to do stuff that is on-par with (and sometimes above) MIT undergraduate classes.

Of course, having done FIRST is not a guarantee of success in 2.007. It is an entirely different ballgame and for the small few of us with FIRST experience, we have to make sure we don't go in overconfident. It will be an interesting experience, though, and I said that I would try to document what it's like to go through it as a FIRST alumni. So, I will update this post over the course of the semester with my progress and my thoughts for the small few who might actually be interested.

POST #1: FIRST look
------------------------
We got a look at the game table today. It's rougly 8'x8', but rounded off on one side. It has several levels separated by parallel ramps on each side, like two roads winding down a mountain. Robots start at the top and work their way down to scoring bins. Bins are worth differing amounts of points based on their difficulty of access, but as far as I know the point values haven't been assigned yet (the table is still being finalized). Their are balls of differing weights (hockey balls and bocce balls) that can be scored in these bins.

The kit of parts (which, not surprisingly, was missing several items) consists of some basic framing metal (1x1 extrusion, plywood, sheet metal, ABS sheets, PVC, aluminum rod. It also has four (small) motors with planetary gears. Power is drawn from two 14.4V drill batteries. Four ~5" plastic wheels and two small casters are provided. Other gears and hardware are available. One important difference from FIRST is that you CANNOT use outside materials with a few exceptions.

Scoring this year for the first time factors in your opponents score: yours + half theirs. (Sound familiar) Also new for this year, the option to use a PIC chip instead of directly wiring motors. This would provide some semi-autonomous (or fully autonomous) function. Usually, motors are directly driven.

Much of the game rules are still being decided and no online documentation is up yet. I will post a link it is, though. For now, though, I still have enough FIRST stuff to do that I can wait for the 2.007 stuff.

[Mike]

I'd be interested in a comparison on how people that have been on a FIRST team do compared to the normal MIT undergrad.

[ZZII 527]

the table:



The small bins are worth more than the large ones. The red bin is worth nothing. The red ball is worth the most, followed by the yellow, then the blue. Final point values and official rules have yet to be established.

[ZZII 527]

WEEK 3:

Full contest rules are now up: http://pergatory.mit.edu/2.007/

My first design review was today. I'm looking to create a versatile drive train that can handle going straight down the angled ledges instead of zig-zagging. I may opt to use the rubber in the kit to make treads. For ball collection, a roller of some type to pull balls up just by driving into them (similar to many FIRST concepts).

Not much movement on the autonomy front. I'm pretty sure even the people running the autonomy workshop don't know what they are doing with it yet...

[ZZII 527]

Sorry if anyone was actually following this - I meant to update it a few more times during the 2.007 build season, but with FIRST season mixed in there too I have been remiss.

WEEK 12
--------------------------------
We are now three weeks away from impound (the 2.007 equivelent of ship date), and the pressure is on to get all modules tested and integrated. The lab is an incredibly busy place these days as people finish all their machining.

I settled pretty early on a tread-driven bot that can climb straight down the ramp sides instead of following a zig-zag path. Since we are only allowed to use the raw materials from the kit, I made the treads out of strips of rubber. At first, there was more friction between the belt and the ground than between the belt and the crowned wheels that hold it in place. The rectify this, I wrapped the outside of the belt in electrical tape (no rule against tape in 2.007). This solved the problem and the result was a very mobile, quick, and robust drive-train than can go from the top level to the bottom level in less than 10 seconds.

I am currently working on the ball collection roller. It is a simple paddle-wheel design than I may convert to something more compliant if it jams to much. I should be finished with all the machining by the end of the week and that will leave me two weeks to tweaking, practicing, and of course non-functionally decorationg.

Here is a CAD render of the current design (minus the treads):


The variety of robot designs is amazing, as the rules are somewhat more lax than FIRST. Things can detach, tether bots are allowed, size and weight limits are not very constraining. Mainly, though, robots fall into a few drive-train categories from what I can see. The most interesting are the ones that stay stationary or minimally mobile on the top shelf, simply grabbing balls and rolling or shooting them to the bins. Done well, this could be a very consistent scoring mechanism and hard to beat. A large handful of other robots are taking the direct path down. Most get over the steps with 4WD and large wheels. Some have treads. The third group are the 2WD robots that will most likely be taking the zig-zag route. Ball collection mechamisms vary greatly too. A lot of passive one-way gates. Several shovels. Active rollers/paddle-wheels make up most of the rest.

Even among the FIRST alumni that I know in the class, there are three very different designs. My tank-inspired stair climber is the simplest. There is a very effective top herder that isolates the balls quickly and drops them down a ramp. The third is an ambitious omni-drive chassis with 6 wheels driven by thee motors, one with a differential. (From only the materials in the kit, this is a very impressive feat. Everything is "home-made.") It may or may not score balls.

I'll try to take some pictures this week as people's designs start coming together.

[ZZII 527]

Final Entry:

After the 13-week build season, the final competition was held this Tuesday and Wednesday night. It was single-elimination format, with some leniency on the first night as there were many double-zero matches. My strategy was simple: grab some small balls, score fast, and block. I survived all the way through to the semi-finals (without ever scoring a large ball) and as a result I will be going to the International Design Competition in Japan in August! The two other FIRST alumni I know of both finished in the top 16 as well (out of a class of 150 or so).

Attached are a picture of my machine, with the red-electrical-tape-coated treads and friction roller (replaced the paddle wheel), and the winning machine, a beautiful design similar in capability to mine but much more effective at picking up the large balls.

[lukevanoort]

Pretty cool, any chance of a video of a match or two showing up?

[ZZII 527]

Originally Posted by lukevanoort

Pretty cool, any chance of a video of a match or two showing up?

There is a nice highlight reel here:

http://web.mit.edu/zigzag/vid/episode6.html
(iTunes video podcast "Zig Zag", Episode #6)

The robot with the scissorjack arm (very beginning of clip), the one with the green lights, and mine (the one with red tank treads) all were made by people who have done FIRST and all finished in the top 16 (out of ~150) .