Odd Autonomous Rules

[Ubiquity]

It looks like there will be a lot of low scoring autonomous periods. I would say it would almost be impossible for a single 'uber' bot score a container set or tote stack in 15 seconds autonomously, but then I didn't think anyone would be able to score three balls high single handed last year. I think it wold be hard to have two robots in teleop grabbing three crates and stacking them in the autozone to do it in 15 seconds. Likewise If you were a container handler with a telescoping arm, even grabbing two containers from the step and moving them would be tough to do in 15 seconds. Ask yourself is it possible to do it in teleop in 15 seconds, before trying to do it in autonomous. The issue is the placement of crates and containers make it hard to get to either quickly.

But ---- you guys are all cleverer then me. I think the safest bet is to make sure you have a good 'move forward 12 feet strategy' that will work from any orientation, and share it with those rookie box bots that can only push a crate on a good day. One of the things that really annoyed me from last year were the number of teams that never even moved at all during autonomous.

[Joseph Smith]

Originally Posted by alopex_rex

In practice I highly doubt that the containers on the step will ever be use in autonomous

What makes you so sure? There are lots of threads discussing the tactical advantage of retreiving the containers from the step as quickly as possible.

[IronicDeadBird]

The q and a said you only get points for one container set

[Munchskull]

Would it be practical to have an autonomous that you program by driving it before hand and recording then having your robot essentialy replay that same drive in autonomous? If so then you could use the practice field to pull on the fly new autonomous programs based on what your team mates can do.

[Christopher149]

Originally Posted by Munchskull

Would it be practical to have an autonomous that you program by driving it before hand and recording then having your robot essentialy replay that same drive in autonomous? If so then you could use the practice field to pull on the fly new autonomous programs based on what your team mates can do.

Technically it should work, but a lot of venues only have partial practice field elements, so you wouldn't have enough field to really try this out.

[Canon reeves]

Originally Posted by Munchskull

Would it be practical to have an autonomous that you program by driving it before hand and recording then having your robot essentialy replay that same drive in autonomous? If so then you could use the practice field to pull on the fly new autonomous programs based on what your team mates can do.

Is this really a thing though? I'm not much of a programmer, but I think it would be pretty cool to have an autonomous that can record and repeat. Of course there could be variables that are different, but is it atleast possible?

[nighterfighter]

Originally Posted by Canon reeves

Is this really a thing though? I'm not much of a programmer, but I think it would be pretty cool to have an autonomous that can record and repeat. Of course there could be variables that are different, but is it atleast possible?

You would need a gyro and encoders, but yes.

Record encoder ticks on each wheel, in combination with the gyro, and have the robot follow that.

It wouldn't be easy, however, and you might be better off just doing it the normal way.

[Munchskull]

Originally Posted by Canon reeves

Is this really a thing though? I'm not much of a programmer, but I think it would be pretty cool to have an autonomous that can record and repeat. Of course there could be variables that are different, but is it atleast possible?

It is totally possible to do. I my self am not a program either but yesterday I brought it up with a mentor and they seemed to be keen on it.

[Canon reeves]

Originally Posted by nighterfighter

You would need a gyro and encoders, but yes.

Record encoder ticks on each wheel, in combination with the gyro, and have the robot follow that.

It wouldn't be easy, however, and you might be better off just doing it the normal way.

Ok, I knew it would involve a lot of sensors, but I could also see a pre programmed autonomous being optimized quicker than a driver could.

[nighterfighter]

Originally Posted by Canon reeves

Ok, I knew it would involve a lot of sensors, but I could also see a pre programmed autonomous being optimized quicker than a driver could.

I guess you *could* attempt it without any sort of feedback, but it wouldn't be accurate or precise.

You would have to record the the values being sent to the Talon/Jag/Victor (I.E. what value the motor.Set() command is recieving), and for how long each motor is receiving that command, then have your code emulate that exactly.

But the problem with doing that without feedback, is if I tell both my drive motors: motor1.set(1) and motor2.set(1) then wait for 1 second before telling them to stop, it will go a different distance every single time. There are a lot of factors that affect how fast/far the wheels will turn...the friction in the wheels, if the wheels are slipping, battery voltage, weight, phase of the moon, what you ate for breakfast...

That's why you need feedback, usually in the form of encoders and a gyro, to tell you how far you have actually gone. (And it might be a good idea to have non-powered follower wheels with encoders attached to them, rather than encoders on the driven wheels. If the driven wheels are slipping (turning but not actually driving the robot), your code will think it has gone farther than it actually has.)

If you want proof of that, write this simple autonomous.

lMotor.Set(1);
rMotor.Set(1);
Wait(2);
lMotor.Set(0);
rMotor.Set(0);

That should drive the robot "straight" for 2 seconds, then stop. (Of course, you might need to make one of those a negative value). Record how far the robot goes, and at what approximate angle. Now do that 5 times.

I'm sure each time will be a slightly different amount.

Now put in a freshly charged battery, and do it 5 times.

You now have 10 different results!

[Munchskull]

Originally Posted by Canon reeves

Ok, I knew it would involve a lot of sensors, but I could also see a pre programmed autonomous being optimized quicker than a driver could.

Depends on what you are doing. Complex three tote auto? Driver ghosting might be easier. Simple drive forward? Programing would be best.

[nighterfighter]

Originally Posted by Munchskull

Depends on what you are doing. Complex three tote auto? Driver ghosting might be easier. Simple drive forward? Programing would be best.

Is 3 tote auto that much more complex than driving forward, though?

Depending on how you pick up the totes, it could be, but it is easy to drive the robot a certain distance, using encoders and a gyro...and we know exactly where the 3 totes are placed.

[BethMo]

Originally Posted by Munchskull

Would it be practical to have an autonomous that you program by driving it before hand and recording then having your robot essentialy replay that same drive in autonomous?

My team did that one year.

[MorrisXX]

Do you need to have the starting recycling containers in the autozone for the points? Or could one of those containers come from the Step?

[BethMo]

Originally Posted by MorrisXX

Do you need to have the starting recycling containers in the autozone for the points? Or could one of those containers come from the Step?

Any 3 (or more) will do. Doesn't matter where they came from.