Fastest Possible Canburglar
 

I wrote a python script to calculate the fastest possible canburglar you could make if it was powered by motors... it appears to be in the 20 ms range. This was inspired by the can burglar from 973 that we were putting on our robot in place of the "tail" at the championship. Let me know what you think.

I attached the python file as canburglar.txt and a pdf which I used for some of the equations. Here is the code:

import math
## Using SI Units - kg, m, s, current in Amperes, angles are printed in degrees but calculations are in radians

##System
t = 0
e = 2.718281828459045
dt = 0.001
rod_mass = .15 # kg
rod_length = 2.5 # m
hook_mass = .05 # kg
J = ((16/48) * rod_mass * (rod_length ** 2)) + hook_mass * (rod_length ** 2) ## rod moment of inertia + hook moment of inertia
angle = 0
terminalangle = 90 * (math.pi / 180) # in radians

#Motor Banebots RS-775 down-regulated to 12v
motor_num = 4
motor_stall_I = 100
motor_stall_T = 1.175
motor_free_omega = 13000 * math.pi * 2
geardown = 1
test_endpoint = 3000
print(motor_free_omega)

def getVelocity():
power_term = -((motor_stall_T * motor_num * geardown)/ (J * (motor_free_omega / geardown))) * t
return (motor_free_omega / geardown) * (1 - e ** power_term)

lowestt = 1, 1, 10000 # low geardown, high geardown, time
while (geardown <= test_endpoint):
#print("terminal velocity = " + str(motor_free_omega / geardown / (2 * math.pi)))
while angle < terminalangle:
angle += getVelocity() * dt
t += dt
print("terminal velocity = " + str(motor_free_omega / geardown / (2 * math.pi)) + " rpm " + " geardown = " + str(geardown) + " angle = " + str(angle / (math.pi / 180) ) + " t = " + str(t) + " velocity = " + str(getVelocity() / (2 * math.pi)) + " rpm " + " \n")
#print("geardown = " + str(geardown) + " angle = " + str(angle / (math.pi / 180) ) + " t = " + str(t))
if (lowestt[2] > t):
lowestt = geardown, geardown, t
elif lowestt[2] == t:
lowestt = lowestt[0], geardown, t

t = 0
angle = 0
geardown += 1

print("fastest canburglar, geardown = " + str(lowestt[0]) + " :")
geardown = lowestt[0]
while angle < terminalangle:
angle += getVelocity() * dt
t += dt
print("terminal velocity = " + str(motor_free_omega / geardown / (2 * math.pi)) + " rpm " + " geardown = " + str(geardown) + " angle = " + str(angle / (math.pi / 180) ) + " t = " + str(t) + " velocity = " + str(getVelocity() / (2 * math.pi)) + " rpm " + " \n")

if (test_endpoint == lowestt[1]):
print("******************************************* ************************************************** **********************************************")
print("**************** Warning, you may not have found the absolute minimum for cangrabber speed. Increase test_endpoint. ***********************")
print("******************************************* ************************************************** **********************************************")
else:
print(" Choose a gear ratio between " + str(lowestt[0]) + " and " + str(lowestt[1]))

Re: Fastest Possible Canburglar
 

Only four rs775s? With four CIMs we got .18 seconds. ;)

Re: Fastest Possible Canburglar
 

What about elastic//spring powered? I could see that being much faster.

Re: Fastest Possible Canburglar
 

You can get slightly faster times if the gear ratio is not constant. A linkage or, non-round pulley or sprocket set can accomplish this.

Do you account for voltage drop? Stalling all those motors drops the voltage considerably.

Do you consider the inertia of the motors? With the gear reduction, they end up making a significant contribution to the inertia of the system.

This is a good first step. Simulations should always start simply, and add other factors and complexity as you prove it out.

Re: Fastest Possible Canburglar
 

This guy knows whats up.

Re: Fastest Possible Canburglar
 

I'll explain when I have a few minutes and not on my cell phone, but we figured out a way to start in the holes legally. It was cheesecaked onto our 4th bot, but since it basically put a robot out of commission it wasn't worth playing in finals. Plus we were not exactly sure if it would be called legal or not, even though it was 100% within the rules and within the field.

Re: Fastest Possible Canburglar
 

I heard about this, double checked all the pertinent rules, and then grinned ear to ear. I'll let you post the details, but I was definitely blown away by your creativity.

One interesting aspect of this game is that the canburglar wars didn't really start until the last day of the season (as predicted by many). A side effect of this is that the canburglar rules/corner cases/how rules would be enforced by inspectors and referees were not really tested until Championships. Harpoon guns, crossbows, tennis balls on strings, hundreds of pounds of spring force, loopholes - there was a little of everything.

Re: Fastest Possible Canburglar
 

I'm interested, posting so I can remember to come back later.

Re: Fastest Possible Canburglar
 

I also found a flaw... hope we've found the same

Re: Fastest Possible Canburglar
 

Your interpretation is entirely legal by the rules, and It definitely would have broken the game. Please post pictures when you can. From the descriptions I can only imagine how massive this thing would be.

The real question is do you think you could do all the set-up required in under 60 seconds?

Re: Fastest Possible Canburglar
 

Ha! Nice find! I wish you had done it, it might have led to some interesting and productive discussions regarding lawyering and the assumed intent of the rules.

I asked a head ref about a similar case, and he said that he couldn't rule without seeing it, but said he would probably have asked the team to come back and fix the offending mechanism.

Re: Fastest Possible Canburglar
 

<G10> does not require set-up in less than 60 seconds, only that set-up must not cause a significant or repeated delay to starting a match. The 60 section portion in the blue box is simply an example. There is not set time to when a 60 second clock would start, how to distinguish it from regular autonomous set up (which frequently takes longer than 60 seconds), etc.

I am very curious as to how it satisfied <G7> B&C while being within the field borders, though.

Re: Fastest Possible Canburglar
 

This was often not enforced as intended. Seeing refs standing around kids with stopwatches at events this year looking to peg people was disappointing to say the least.

Re: Fastest Possible Canburglar
 

FIELD includes the width of the guardrail, and LANDFILL ZONE doesn't include the "and including" wording. That's my guess.

Re: Fastest Possible Canburglar
 

Bingo

Re: Fastest Possible Canburglar
 

The rules state that the landfill is bounded by but not including the guardrail. However the rules also say that the field is bounded by and including the guardrail. And since G7 just states you can't be in the landfill, and must be fully supported by the carpet or scoring platform, it is legal to actually go around the landfill. And since the guardrail is said to be inside the field, it becomes really hard to actually call it illegal. Now mechanically this device is very difficult to make, because it has to be cantilevered 6 ft out and then 10 feet into the field. We tried building it out of square tube, and twisted the bar close to 70 degrees. But in the end we did have a 2 can version working.

And yes it was able to be setup in 60s. Their drive team was practicing all through elims.

This game actually had a ton of loopholes. Like for instance purposely taking a G7 penalty was still only a disable. So technically if a team didn't mind getting a robot disabled, they could have started in the holes, and had another robot pulling that bot, and potentially using the manual actuation buttons on the pneumatics to move arms and such.

I just wish we would have figured this out before last week so we would actually have had time to build something cool.

Re: Fastest Possible Canburglar
 

So the definitions in the rules on pg 9 I believe are much clearer then the glossary. And there was a Q&A stating that the rules overrule the glossary.

Re: Fastest Possible Canburglar
 

Hmm, I was thinking in a different direction: G7 states that when the robots are placed on the field they must be outside the landfill. It doesn't say that the robot has to stay outside the landfill there until the start of the match. So have a slow-moving mechanism that slowly flops down to the can during the time before the match starts, and while the FTAs are trying to fix the broken FMS system.
My loophole is much more likely to be closed immediately by a ref. A ref actually told me that if our canburglar was to be spring-loaded and accidentally move into the land-fill he would likely ask us to fix it, or disable us if it happens again.

You should have tried carbon fiber tube to build your mechanism, but that is a lot of cash to drop on something that could be decided to be illegal.

Re: Fastest Possible Canburglar
 

It was too bad we couldn't put out "Jenny" (our internal code name for the project) at the event. We were all very eager to see reactions when we pulled out a 0 second step can grabber haha

If this game is played at IRI, I look forward to seeing some really crazy solutions for canburgling. Hopefully we'll see some harpoons in action!

Re: Fastest Possible Canburglar
 

Okay admittedly I am new to the nuances of creative rule interpretation as a first year mentor, but I don't see it. The Landfill is described in 2.1.5 as "an infinitely tall volume ...". The term volume implies bounded in 3 dimensions so I do not see how going in from the end, if I am understanding your plan, qualifies as being "outside'" the landfill as required by G7c. In my world, there is no way to be in contact the RC without being inside the defined volume.

Re: Fastest Possible Canburglar
 

One volume is bound by the outside of the guardrail, and the other is bound by the inside, by the letter of the rules. There is technically a guardrail-wide channel in which the robot can snake a structure through to the step without ever entering the landfill volume.

Re: Fastest Possible Canburglar
 

Does this help?

Re: Fastest Possible Canburglar
 

Yup! We implemented just one side (2 cans) due to the tremendous torque on the system with the circular aluminum tubing we could source before worlds. If our team had the time, budget, and experience with carbon fiber we would likely have built an attachment to go all the way across.

Karthik always says "Read the rules" and our team has taken that to heart (I love his talks. If you haven't seen them you should). :D

Re: Fastest Possible Canburglar
 

I was definitely surprised by the amount of different mechanisms for grabbing the trashcans that came around after the last day of regionals. I don't know if the judges really expected that onslaught of teams trying to grab the cans in so many different ways and I have no idea what was up to inspection. The coolest thing, I would say, about the canburgler wars at worlds was the fact that it was really the only "defense" in the game this year so it was interesting at certain points to see that play out.

Re: Fastest Possible Canburglar
 

The only real hindrance would be the speed of your release mechanism. I was imagining a 2-speed gearbox with a set of gears taken out. Use the one remaining ratio to draw back and then shift to where the second ratio would be. #pneumaticquickrealeaseftw

Re: Fastest Possible Canburglar
 

We used a crossbow type trigger pulled by a motor. We could only fire once, then we used 2 pistons to release the surgical tubing bundles and had it tensioned to get pulled back up(which it fired against)

Re: Fastest Possible Canburglar
 

The real advantage to this system is it could be designed to block any other burglar attempt because it was already there, not necessarily that it would be pulled back so quickly. Shield the hole, shield the handles, and pull back in a reasonable amount of time and that thing could be unbeatable.

-Brando

Re: Fastest Possible Canburglar
 

And it was :)

Re: Fastest Possible Canburglar
 

I heard a rumor that Karthik mentioned de-ionized water and freezing motors in his Wednesday talk w.r.t. speeding up canburglars. Can anyone explain how this helps and what the process of it is? I vaguely remember something from the Minibot days as well, but it's a bit fuzzy.

Re: Fastest Possible Canburglar
 

Likely trying to decrease the resistance of the internal wiring of the motor? Think super conductor.

Re: Fastest Possible Canburglar
 

Sadly I never saw you guys testing this, do you guys mind posting a picture?

Great playing with you on Carson... we should have had it :)

Re: Fastest Possible Canburglar
 

In the rush of things I'm afraid we didn't take any pictures. Once we get our pit and crate back we can take some pictures and video to show everyone! :)

Re: Fastest Possible Canburglar
 

I actually mentioned it with respect to people trying to speed up minibots. However, I would not be surprised if teams tried this for Canburglars this season.

Re: Fastest Possible Canburglar
 

Last year, I worked up a linkage-based design for this purpose to power a yoga-ball catapult with motors. We didn't use it, but I recreated it and posted a spreadsheet describing it recently on the "Launch Unusual Items" thread . I also did a cardiod pulley based design, but that's probably not too useful in this context, as you want to be able to pull the arm back up later in the match.


And yes, springs can be much faster, but making them both safe and reliable is probably beyond most teams' capabilities.

Re: Fastest Possible Canburglar
 

Just curious - with the minibots, I haven't dug for it, but none of the descriptions I've seen of the minibot tricks involved raising the volume of the minibot faster than its center of gravity. It seems to me that the main part of the minibot could be (for example) half the maximum height, standing on lightweight pass-through stilts that put it at the maximum height. This would put the CoG at about 70% of the minibot height. Then, as the minibot climbs, the stilts are lifted relative to the COG (probably geared right into the same motor) until, at the end, the main part of the minibot is occupying the bottom half of the volume, so the CoG is at 30% of the minibot height. By decreasing the amount of work to be done, you should be able to execute faster.