3310 Robot Reveal

[Paul Copioli]

Team 3310 presents "Warhawk".


Features:

Auton Bin Grabber to acquire and move 2 bins off the step in under 0.25 seconds.

4 degree of freedom "Palletizing" arm designed to park at the Chute Door (yes, Chute Door) and stack up to 5 columns of totes up to 6 high.

Arm cycle time for a round trip (chute door to tote placement and back to chute door) is approximately 3.5 seconds.

Link to teaser video: http://youtu.be/pmyphoi-ASg

Needless to say, we are excited for Dallas.

[AllenGregoryIV]

Been hearing rumors all season about this thing. Glad it's out in the wild. Can't wait to see this machine in action. Hopefully we'll get to play together at Bayou.

The quarter sec can grab is insane.

[Jaywalker1711]

How are those arms so fast?!?

Well done!

[Paul Copioli]

How are those arms so fast?!?

Motion profile + a ridiculously oversized motor for the job + fishing pole whip

Basically, the brute force method.

[Foster]

Been hearing rumors all season about this thing. Glad it's out in the wild. The quarter sec can grab is insane.

Not to appear to be silly, but I think they got the idea after your user photo. Isn't that what the arm on the left is doing?? :rolleyes:

[Scott Morgan]

That's one fast can grab.

I'd like to see this "Palletizing" arm in action. Looks way different than the other stack building systems I've seen so far.

[AdamHeard]

I had been joking that I was hoping you'd make a FANUC arm when you came back.... you did it.

Very excited to see it!

[s_forbes]

4 degree of freedom "Palletizing" arm designed to park at the Chute Door (yes, Chute Door) and stack up to 5 columns of totes up to 6 high.

Arm cycle time for a round trip (chute door to tote placement and back to chute door) is approximately 3.5 seconds.


I got really excited when I read this and saw the youtube link, but alas... The blazing fast whips are still cool.

[mman1506]

Is the arm mounted on a turret?

[JohnSchneider]

Is the arm mounted on a turret?

Yes. 360 degree turret is joint 1, large arm is joint 2 smaller arm is joint 3 and wrist is link 4.

[Addison4300]

I like the look of the black metal!

[Ty Tremblay]

Yes. 360 degree turret is joint 1, large arm is joint 2 smaller arm is joint 3 and wrist is link 4.

4 DOF! What kind of controller are you using? Are you calculating the forward and/or inverse kinematics?

[Andrew Lawrence]

Looks beautiful. And that .25 second bin grab is very impressive. Any chance we can see the stacking arm in action?

[Paul Copioli]

4 DOF! What kind of controller are you using? Are you calculating the forward and/or inverse kinematics?

Doing both forward and inverse kinematics, when required.

Using the RoboRio for kinematics, inverse kinematics, coordinated motion and motion profile; using the Talon SRX for all PID and sensor data collection.

Roborio simply sends position commands to the 4 Talons every 10ms.

[Ty Tremblay]

Doing both forward and inverse kinematics, when required.

Using the RoboRio for kinematics, inverse kinematics, coordinated motion and motion profile; using the Talon SRX for all PID and sensor data collection.

Roborio simply sends position commands to the 4 Talons every 10ms.

That's excellent. I love how the SRXs are enabling FRC teams to use distributed control. Are you updating the PID constants dynamically or did you find a set for each joint that works?

[Paul Copioli]

That's excellent. I love how the SRXs are enabling FRC teams to use distributed control. Are you updating the PID constants dynamically or did you find a set for each joint that works?

Gravity loaded axes are dynamic due to the differences between with tote and without, but J1 and J4 are static constants.

[Tom Bottiglieri]

Motion profile on the can grabber? So you don't bounce off the top?

[Kyler Hagler]

Motion profile on the can grabber? So you don't bounce off the top?

Motion profile for the 4 axis arm. Can grabbers are encoder controlled via motion profile also.

[Lil' Lavery]

Really want to see that arm in action. Hopefully I can make it to St. Louis this year to witness it first hand. :cool:

[Paul Copioli]

Motion profile on the can grabber? So you don't bounce off the top?

Correct.

[wireties]

Beautiful machine guys - congrats!

[tomead]

Beautiful machine guys - congrats!

Where is yours?

[Aren_Hill]

Where is yours?

In some bags.

[wireties]

our videographer is on strike!

[lynca]

I'm looking forward to see how fast this thing operates !

It might be the 1st industrial ARM in FRC. :eek:

[Scott Morgan]

I'm looking forward to see how fast this thing operates !

It might be the 1st industrial ARM in FRC. :eek:

I seem to remember on that was kinda similar in 2005, I don't remember the team off the top of my head, but I think I have a picture at home.

[Mark Sheridan]

wow, i guess the middle RC arms race is upon us.

I can't wait to see the arm in action. Very cool robot.

[BBray_T1296]

wow, i guess the middle RC arms race is upon us.

I see what you did there. :D

[asid61]

Looks fantastic! I thought pneumatics would be faster, but I guess motors do the job just as well!

[Jared Russell]

wow, i guess the middle RC arms race is upon us.

Yes, because who doesn't want to re-live minibots all over again?

[Thad House]

Yes, because who doesn't want to re-live minibots all over again?

And that's my problem with this years game. Just instead of being won in the last second, it gets won in the first second...

[ozrien]

That's excellent. I love how the SRXs are enabling FRC teams to use distributed control...
+1 :)
It's a new era of robots...

[JohnSchneider]

And that's my problem with this years game. Just instead of being won in the last second, it gets won in the first second...

First 1/4 second ;)

[hrench]

You've invented the robot that can play any game --frisbees, basketballs, etc, with just a change of manipulator.

[Lil' Lavery]

And that's my problem with this years game. Just instead of being won in the last second, it gets won in the first second...

You still have to outscore the other alliance after getting those center containers. Scoring 301* points every match is still going to be hard, even with 7 recycling containers.


*I leave the version where you reduce their score further as an exercise to the reader

[JVN]

Another amazing robot from 3310! While everyone on 148 is impressed with your can grabbing, we're very disappointed that this year's robot isn't just 4 crossbows and a giant winch.

See you guys next week. :) #TeamTexas

[Mark Sheridan]

Yes, because who doesn't want to re-live minibots all over again?

and when two robots tie, they get to re-live 2002. the nostalgia is too much.

[Kyler Hagler]

Another amazing robot from 3310! While everyone on 148 is impressed with your can grabbing, we're very disappointed that this year's robot isn't just 4 crossbows and a giant winch.

See you guys next week. :) #TeamTexas

We did contemplate a grappling hook system with 4 crossbows but then we decided on something that sounds scarier when executed. Plus we wanted to do something we haven't done yet. KARATE CHOP ;)

[Thad House]

You still have to outscore the other alliance after getting those center containers. Scoring 301 points every match is still going to be hard, even with 7 recycling containers.

Getting 3 full stacks with cans and litter, and a perfect auto is only 158. This would be for the team that loses the cans. 5 full stacks with cans and litter is 210, assuming no auto points. for the teams that win cans. The only points the alliance with 3 cans can get is from litter across the field and individual totes. Assuming 7 pieces of litter (28 pts) across the field, they would now have to score another 12 totes just to even catch up with 5 stacks with cans and litter. That might not look like alot, but every piece of litter missed, or put into the landfill zone increases that tote number by 2. It could get out of control real fast, especially because at the high level, I expect the captain and #1 pick to both be able to do at least 3 full stacks with cans. 6 stacks with cans and litter is basically insurmountable, because it would require 21 totes to be scored to make up for that stack.

[Lil' Lavery]

Getting 3 full stacks with cans and litter, and a perfect auto is only 158. 5 full stacks with cans and litter is 210, assuming no auto points. Since if you control 7 cans, the only points they can get is from litter across the field and individual totes. Assuming 7 pieces of litter (28 pts) across the field, they would now have to score another 12 totes just to even catch up with 5 stacks with cans and litter. That might not look like alot, but every piece of litter missed, or put into the landfill zone increases that tote number by 2. It could get out of control real fast, especially because at the high level, I expect the captain and #1 pick to both be able to do at least 3 full stacks with cans.

Sounds like a lot of stuff you have to do after the first 0.25 seconds to win the match, to me. ;)

An alliance can still score 300 points without any step cans.

[Thad House]

Sounds like a lot of stuff you have to do after the first 0.25 seconds to win the match, to me. ;)

An alliance can still score 300 points without any step cans.

Technically yes, if they can score 57 totes in addition to their 3 stacks. You'd have to be pretty creative, because thats 10 stacks. That makes 13 stacks total, and if an alliance can get 13 full stacks to actually fit on the scoring platform I would be very impressed.

And yes its not auto win, but neither were minibots. You still had to put tubes up to win. Its just that you didnt have to do a perfect rack, unlike what the other team would have had to do.

[Lil' Lavery]

Only 52 "additional" totes (assuming one recycling container per stack).

[asid61]

Dallas is going to be really unfair to the good teams. It's unlikely that all these amazing teams will be on the winning alliance.
On the other hand, that's pretty much every year, isn't it?

[cglrcng]

But, being assured (anyone being able to pull that off consistantly within 1/4of a second), that you have at least an even amount of RC's, assures an even chance at max points for your alliance. EQUALS VERY IMPORTANT.

Great job there Team # 3310. A great robot.

[Lil' Lavery]

But, being assured (anyone being able to pull that off consistantly within 1/4of a second), that you have at least an even amount of RC's, assures an even chance at max points for your alliance. EQUALS VERY IMPORTANT.

Great job there Team # 3310. A great robot.

Not if the other alliance gets more of the totes from the step than you do, or is more efficient with their litter scores.

[asid61]

But, being assured (anyone being able to pull that off consistantly within 1/4of a second), that you have at least an even amount of RC's, assures an even chance at max points for your alliance. EQUALS VERY IMPORTANT.

Great job there Team # 3310. A great robot.

But there are multiple teams with a 0.25 second grab (some that haven't done a release video, I'm certain) so if two end up on the same alliance, bad things will happen.
It's also totally possible to have a 0.25 second 4-can auton with large enough pistons.

[AdamHeard]

But there are multiple teams with a 0.25 second grab (some that haven't done a release video, I'm certain) so if two end up on the same alliance, bad things will happen.
It's also totally possible to have a 0.25 second 4-can auton with large enough pistons.

.25 seconds is well near where it's hard to go that fast without intelligent control. Hence the motor on 3310.

The faster you go the more energy you need to dissipate to stop.

[magnets]

.25 seconds is well near where it's hard to go that fast without intelligent control. Hence the motor on 3310.

The faster you go the more energy you need to dissipate to stop.

Why not have the container dissipate the energy? The sort of hook 3310 and 118 have might bounce out, but a latching hook wouldn't have this issue. Also, if you hit the handle hard enough, you'd move the container out of the way to prevent anybody from grabbing it without having to worry about timing to drive the robot back.

I just can't wait to watch some tug of war!

[s_forbes]

The faster you go the more energy you need to dissipate to stop.

Why do you need to stop?

[JohnSchneider]

Why not have the container dissipate the energy? The sort of hook 3310 and 118 have might bounce out, but a latching hook wouldn't have this issue. Also, if you hit the handle hard enough, you'd move the container out of the way to prevent anybody from grabbing it without having to worry about timing to drive the robot back.

I just can't wait to watch some tug of war!

We lock into the top of the bin.

[Selle2.0]

Using tractable hooks, like a harpoon. Once we have it, it's not coming out.

[Tom Bottiglieri]

.25 seconds is well near where it's hard to go that fast without intelligent control. Hence the motor on 3310.

The faster you go the more energy you need to dissipate to stop.

That's still only 25 control packets :ahh:

[WillRobinson]

What happens if a 4RC grabber can latch into the top of the RC equally as fast but then 3310 takes off? Wouldn't 3310 likely pull one end of the 4RC grabber across the step?

Would be an interesting tug of war to say the least.

[JamesTerm]

Using tractable hooks, like a harpoon. Once we have it, it's not coming out.

What happens if a 4RC grabber can latch into the top of the RC equally as fast but then 3310 takes off? Wouldn't 3310 likely pull one end of the 4RC grabber across the step?

Would be an interesting tug of war to say the least.

Indeed... ever hear the story of how hunters capture monkeys where they get their hand stuck in a jar and will not let go?

You may wish consider letting go as a failsafe, and have a backup plan.

[Paul Copioli]

Indeed... ever hear the story of how hunters capture monkeys where they get their hand stuck in a jar and will not let go?

You may wish consider letting go as a failsafe, and have a backup plan.

James,

We can let go ... if we decide to:)

[Cash4587]

I could be wrong but it seems like yall are sitting in the landfill to grab the cans? I didn't think you could just flop an arm down onto a can with your robot sitting stationary in auto and still be within the height limit.

[JohnSchneider]

I could be wrong but it seems like yall are sitting in the landfill to grab the cans? I didn't think you could just flop an arm down onto a can with your robot sitting stationary in auto and still be within the height limit.

No, we start fully out of the landfill zone. And geometry is your friend :)

[Wayne TenBrink]

When do I get to see video of the stacking arm?

[JamesTerm]

James,

We can let go ... if we decide to:)

So do the harpoon hooks retract to let go... or do you let go from the other end and lose the arms... probably be a good idea to share how to let go... as that would be a benefit to everyone... especially to the person who still wants to hang on.

[cjl2625]

No, we start fully out of the landfill zone. And geometry is your friend :)

So do you move forward into the landfill, or does your arm reach from the auto zone?

[TheOtherGuy]

Paul, I'm hoping 3310 is going to release at least the controls code for the arm at some point. I think the FRC community would benefit greatly from a FANUC view of arm control ;)

Anyway, fantastic robot! My second favorite this season. I'll be watching Dallas closely.

[PAR_WIG1350]

So do you move forward into the landfill, or does your arm reach from the auto zone?

Considering you can't start in the AUTO ZONE or the LANDFILL ZONE, I imagine they actually reach from the space that exists between the two zones, assuming they don't move.

G7 When placed on the FIELD for a MATCH, each ROBOT must be:
A. in compliance with all ROBOT rules, i.e. it has passed Inspection. For exceptions regarding Practice MATCHES, see
Section 5.2 – Practice MATCHES.
B. fully supported by the floor, SCORING PLATFORM, and/or SCORING PLATFORM ramps on their ALLIANCE’S side of the
FIELD, and
C. completely outside of their AUTO ZONE and LANDFILL ZONE

[NWChen]

Seeing 3310's smooth movements on the Dallas stream blew my mind. What sort of device/interface operates the palletizing arm on the driver station?

[Paul Copioli]

It's autonomous ... mostly.

The arm operator initiates the cycle. He can pause and resume if the gripper is off a little bit. Each cycle the robot pauses and waits for the ok from the arm operator. If the gripper is in the right spot then he hits go. If not then he can pause and manually adjust.

We are still working on speeding it up and getting our robot to the right spot in the first place. Should be ok today.

[Alex Cormier]

I had been joking that I was hoping you'd make a FANUC arm when you came back.... you did it.

Very excited to see it!

I noticed that Paul was on a different team and saw this robot on webcast. Knew exactly where it came from, it's just black not yellow. :p

[Ty Tremblay]

Any video of it in action? I can't watch the stream today.

[Akash Rastogi]

This thing is so fun to watch. Great job!

[Andrew Lawrence]

This thing is so fun to watch. Great job!

I made the mistake of blinking and missed their bin grab.