Things to Explore

[apalrd]

Originally Posted by davidthefat

...so there is no "front". Now that is a mind blower...

You can pick which side your front shall be. A robot has no front, and is only limited by its directions of travel, usually this lends to having two directions of motion which are labeled "front" and "back", but really it makes no difference.

For example, we have a robot (from 2006) which has two "fronts". You use one "front" to collect the balls, then press a button to change the "front" and shoot out of a different direction (Which one might call the "back" while collecting), but really it has two fronts. It does not matter in code.



But seriously, if you've never built a 6wd, it is a start.
If you seriously want to, a swerve can be advantageous (if you can learn to drive it, something most teams overlook), but it is very hard to build so I would not recommend it to a team who has not yet mastered a 6wd.

Conveyors and such: We generally use orange belts (I have no idea where they come from), and drive them with rollers. Sometimes we use flat, sometimes we use round. Ball manipulation is sometimes important to the game (09 and 06, for example), and it might be helpful to prototype, but it isn't very complex (Mostly spinning drums, not much complex motion).

Linear racks and pinions: I haven't seen these used a whole lot. Most linear systems use a drum and cable system, and those can be very tricky. If you like using linear elevator systems, this could be prototyped.

Manipulators: History has shown that classic roller claws work very well. Other options include pneumatic claws. These two are the most generic manipulators for handling odd game pieces.

Arms: These are fun. You get to do all kind of load and gear ratio calculations, then program it to not kill itself, be as fast as possible, have a little error as possible, and be awesome. These each require specific implementations in software, such as a basic PD/PID loop (You will almost certainly need D, and probably I too), and if you have multiple segments you can correct for error on a joint by compensating in the successive joints. Since the cRio has a lot of power, you can do realtime trig to determine the angle for each segment, then recalculate based on where the joints actually end up, correcting for error while keeping the target end point, for example, above a certain height. If you want to, you can also slide the end effector in and out while keeping it at a fixed height, and determine all of the angles on the fly.
See if you can design it to be as light as possible, remembering material selection and thickness (For example, .030 aluminum tube is a lot lighter than 1/8" aluminum box, and a whole lot lighter than 80/20).

I don't think line tracking is useful. I have never seen it used successfully, and generally encoders and gyros can get you close enough.

Target tracking: See if you can do it without overloading the processor. Anyone can do it. It's the great that can do it, while doing something else at the same time.

Pneumatics: Use them like any other actuator. They have two positions, if you need something in two positions (like an AndyMark shifter, for example), use pneumatics.


Things to add to the list:
New chassis designs. Do something you haven't done before. Learn something. Build stuff. Have fun.




@davidthefat:
No matter how much brains you have, it doesn't matter if you can't move.