Proposal: New Extension Rules to Fix Contact in Frame Perimeter Damage

Competition Rules/Strategy
1

I have great respect for the GDC, and really enjoyed Rapid React. That said, G204/G205 was a bit of disaster this year, as predicted. Not only did it add an unpredictable number of fouls to almost every match, it completely failed to prevent robots from damaging each other, sometimes violently.

So what can we do in 2023? I’ve had some ideas, but none of them stand up to much scrutiny:

  • The “Do Nothing” option is untenable. G204 seems reasonable on paper (stay out of other robots) but in practice, extensions + a game like rapid react is always going to have contact in the frame perimeter. It’s unfair to give teams the impression that they’ll be able to build a robot that extends without incurring fouls (see the 2022 Everybot intake).
  • Previous versions of the rules (where contact had to be "damaging) are also untenable. By the time contact is “damaging” it’s too late, and it’s unfair to expect refs to judge that during the heat of matches.
  • Expecting contact, and requiring robust construction (or mandatory perimeter guards or something) is untenable because it is too much to ask of weak teams and it can become an arms race. Plus, what standards are we supposed to use for “robust”.

Instead here is a proposal for a future “Robot Contact” rule:

  1. No robot may extend more that 6" outside of it’s frame perimeter during teleop (in shared zones).
    This limit will keep mechanisms within the “bumper zone” between robots. Nothing should be able to reach into another robots frame perimeter.

Obviously the game will have to be designed around this rule. Game pieces / field elements will have to be sized/shaped so that they can be manipulated with only 3" of reach beyond the bumper in shared zones. The GDC can set up “safe zones” or gameplay periods (including endgame) where larger extensions are allowed. And robots will likely end up with more bumper cutouts. But those are pretty minor accommodations.

I’m interested to hear what others think the pros and cons of this rule is, and what adjustments additions should be made. I hope at minimum we can agree that something should change for next year. We can’t keep going on pretending that interior robot damage shouldn’t happen, and then shrugging when it does.

Thanks to @AllenGregoryIV for the suggestion and discussion.

14 Likes
2

My only problem is pool noodle robustness. After a season of playing and being abused, I’ve seen noodles unintentionally get squished significantly, enough to enable a rule like this to still allow contact inside the frame perimeter. That’s not to say it’s a bad rule, just that it may not completely solve the problem.

Back in 2010, there was a rule about the length of time a mechanism could be extended - something similar could lead to more active control over mechanisms, and briefer extensions, but does pose a larger hurdle for inexperienced teams / team members.

Another thought: all extensions have to happen entirely within the bumper zone. Any contact inside the frame perimeter would then, by definition, have to be inside a bumper opening, a risk that teams can more easily plan for.

7 Likes
3

I personally think that the fact that the cargo were so large this year is what made for so many issues. In the 2022 game, teams often made intakes that would go above other teams bumpers, making the most common collision(excluding bumper to bumper) this year one teams intake that goes over another teams bumpers. With the 2020/21 game, since the power cells had a smaller diameter than the cargo, more intakes would be running into opponents bumpers than over them.

I know that the GDC has been very static on their bumper rules, but maybe they could look into changing them for certain games. Especially during a shooting game with many over the bumper intakes and a game piece that would require the bottom of the intake to be above the top of the bumpers, this seems like an obvious addition. Having the bumpers not protect against a mechanism that is most likely to be out of a robot’s frame perimeter seems like an oversight.

I know that this would mean that teams could not reuse bumpers through years, but this small drawback seems worth it when weighing what could happen otherwise.

3 Likes
4

So extension rules won’t solve every issue for example:
If you watch 233 and 179 interact in this clip you will see no frame extension, no one moving at high speeds. Yet in a way I can only describe as “somehow” 179 just pops up into the 233 robot. This interaction resulted in 233 having their battery disconnected.

Removing the extensions would not have prevented this and I am unsure how I would have called this as a ref since this is bumper to bumper contact and then someones “bumper zone” interacting with someone elses not “bumper zone”

9 Likes
5

I wouldn’t want to see more restrictive expansion limits, but I have given a lot of thought to G204/G205. Here’s what I’m starting to settle on:

Remove G204 entirely. Inconsequential contact should not be penalized.

Reword G205 to be a card for any damage which affects robot functionality, defined as the ability to navigate the field or score points. Yellow for offenses that did not affect the outcome of the match, Red for offenses which did. Implement video review exclusively for G205. It is my understanding that the streaming A/V equipment is now shipped with the field hardware and is basically standardized across events. Integrate that infrastructure into the FMS so it is easy for a head ref to review a match immediately after a team shows them the damage to their robot, to ensure it meets the criteria for a G205 infraction.

Source: a drive coach for a team that received a lot of G204s while strictly playing offense, half of which were called correctly according to the initiation-of-contact clause. Also having lost eliminations matches on legitimate G205 calls as well as missed G205 calls.

8 Likes
6

With respect, I disagree. With extensions teaching into robots, there’s no way for drivers to control the “consequence” of contact, no way for builders to prevent it, and no reasonable way for refs to assess it, even with video review. It’s just a recipe for tears.

2 Likes
7

As much as I agree that this is the ideal solution we are not there yet. FIRST Webcast Unit is only shipped with regional/HQ fields, used by only some regions while many districts use their own solutions, plus there is currently zero integration between FIRST Webcast Unit and the FMS outside of the Audience Display.

7 Likes
8

This has been a problem for some time. Our last 2 robots have had side shields to prevent intrusion. We use 3mm or 6mm birch plywood with carbon fiber laminated to it. Very tough. It is also used to mount electronics and has been used as a structural mounting point. At district champs we took a high speed impact to the side that would have caused serious damage. The shield had a 3/4" puncture but no intrusion. After a season it is fairly scraped up but still good. We mount our air tanks to it to prevent the IRI type explosion. The shields do make access to some parts of the robot difficult. I’ll take this over a shredded damaged robot. Our though is this frame intrusions will always be a problem. Deal with it. No rule change will stop it. Maybe teams should just armor their robots.

8 Likes
9

I think this is the biggest problem with this idea. With every game piece in recent games except for the panels in 2019 you would have needed a bumper/frame cutout to manipulate the game piece. I don’t think that forcing teams to incorporate frame and bumper cutouts is a good idea.

A lot of teams (yours included) designed their 2022 intake to be flexible such that the possibility of the intake being damaged or damaging the other robot in the event of a collision was reduced. Using a small diameter pneumatic cylinder to deploy the intake also causes the intake to retract with only a small amount of contact force. But I think requiring a flexible extension or an extension that can be pushed back into the frame perimeter with a small force would also be to restrictive for less experienced teams to implement.

I think that a rule that would require any items that could be damaged be located behind a plywood or lexan shield (which could include the bumpers) would protect those items allowing robot to robot interaction with minimal risk of damage. The height of those shields could be limited to some height above the floor and then extensions outside the frame perimeter could also be limited to that same height (or less) except for during the end game or in certain zones on the field (protected zones).

This would put the onus on teams to protect their own robot from the potential of robot to robot interaction. You could exclude such shields from the robot weight limit to allow teams to use robust plywood panels if that is what they want to use. You could also allow these panels to be mounted to the outside of the frame perimeter.

3 Likes
10

A 6in extension could still destroy air tanks in a collision like this

image

With an air tank mounted such as this

You could just ban extensions outside the frame perimeter all together.

5 Likes
11

I agree in every way. Both times this year when our team was disabled/damaged by contact, it was not by an extended mechanism. Both times it was by a robot riding up over our bumpers and the contact was bumper/frame to interior robot parts. As you say, a no extension or limited extension rule would not have prevented either of these situations.

2 Likes
12

What product did you use to laminate the birch?

13

Part of the issue is the drive train arms race. Remember 2009? GDC wisely said no extensions since the robots would sliding into each other. So basically you are advocating for no practical extensions because robots are going to hit each other. We are now to the point that some drive trains are so fast and swervy drivers cannot react fast enough to avoid collisions. In fact, I’d say we have the potential for harder collisions than 2009.

So is the issue safety? Unintended damage? Or unintended foul points?

In my opinion, it is too early to tell if the current G204 needs to be changed. I believe we need a stable rule set to help us design robots with foul/damage risks in mind plus it would help the refs to call the rules more consistently.

I thought G204 would be called more like basketball block / charge. Did someone sitting still with an intake out “initiate” contact? As it turn out, the answer was yes. Any contact from an extention was “initiated.” Seems very easy to call, much easier than a block / charge type decision.

Even with that, there were only 2.4 average foul points per match which was 4.6% of the average score. This is fairly typical percentages from recent years. There was no foul point disaster in comparison. (Fouls in 2020 were 11% of the score.)

From a design standpoint, if extension = foul risk then teams will come up with their solution to that instead of forcing the GDC to effectively eliminate extensions as the only design option.

I say keep G204 as is.

David

6 Likes
14

I applaud the effort, as it is an issue. Nonetheless, limiting extensions does not solve the problem.

One action could be to make clear to teams that causing damage is a serious offense. Drivers be warned! Penalties could be commensurate and significant as a deterrent. (Drivers already know how to avoid contact in protected zones, right?).

Another would be to remind teams that such actions, although prohibited, can and do occur. The astute student would consider this for their design. I suspect a limited effectiveness though, as I’d guess ‘everyone already knows this’.

A rule always can be lawyered. Referees can always miss something. It is a difficult problem, and I’m not smart enough to solve it.

2 Likes
15

Determining if damage affected the outcome of the match can be extremely difficult and getting it wrong can end teams’ seasons and leave everyone unsatisfied – similar to the current rule. It also depends on which specific robots received damage. A top end scoring alliance captain having their intake break is much more likely going to impact the result than the 24th team at the event who is just playing defense. I don’t think refs should be responsible for knowing the capabilities of each robot and alliance strategies.

7 Likes
16

Every year we find something on the robot to use carbon fiber lay ups. Our E panel belly pan is birch plywood laminated on both sides with 5.7 OZ plain weaver carbon fiber cloth. My goal is to have every student in our program have basic experience with composite manufacturing methods. Our side panels are made the same as the belly pan.

17

2022’s rise in inside-bumper contact is the result of two factors:

  1. The evolution of the FRC design meta
  2. Game piece selection

Pick-and-place games have always had a certain degree of risk of collisions involving extended arms or game piece manipulators. But teams could also (typically) retract these devices if they weren’t near where they loaded or scored their game pieces, so contact was often easier to assess than with intake devices. Not to say there weren’t controversies (there was definitely a fair bit of heartburn involved in defending and fighting for position around the rack in 2007), but they were generally more isolated and/or egregious events rather than typical gameplay. And eventually the implementation of protected loading and/or scoring areas (and extension limits on defensive bots in 2019) made it increasingly rare for teams to be driving around opposing teams with extended mechanisms that posed a risk of inter-frame contact. The issue these days are with shooting games and the “over bumper intakes” teams use to collect game pieces.

For most of FRC history, “over bumper intakes” were rare. Bumpers didn’t exist at all until 2006 (2005?) and weren’t mandatory until 2008. For ball-games, teams generally loaded balls thru a gap in their frame. 2006, 2009, 2010, and 2017 didn’t even allow for extensions beyond the frame perimeter (and 2010 also didn’t allow for fully intaking a ball and moved the bumper zone to a heigh above the ball). 2008 saw some of the first “over bumper intakes,” but it was far from universal given the bumper rules at the time permitted teams to essentially leave one side of their robot (often the front) without a bumper, and the extension rules that year (you could not extend by a 84" diameter vertical cylinder) limited potential interactions with that truly massive game piece. It wasn’t until 2012 that we really saw the first wave of “over bumper intakes” in action, with teams like 341, 254, and 2056 popularizing the design almost by mistake, as many of the teams who used the design arrived at it while seeking out ways to manipulate the balancing bridges that year. And even still plenty of high end teams (like 1114, 233, and 118) were still loading balls thru a frame gap in 2012.

2014 pushed over-bumper intakes to become the norm. Bumper rules were more restrictive than in 2008 (the last time there was a huge ball game piece). And it was also a game in which extensions beyond the frame perimeter became a big center of attention and controversy regarding fouls (far more so than even 2022, given that tech fouls were worth 50% of an average pre-CMP match score). 2016 saw a reduction in both over-bumper-intakes and also frame perimeter intrusions due to the fact teams could only collect one (much smaller) ball at a time, and the protected zones created by the outer works and trenches.

In 2020 and 2022, over-the-bumper intakes became the de-facto choice for intakes, with far more teams using these than alternative intake styles. So why was there so much more heartburn about intakes in 2022 than 2020? (Hint: it has nothing to do with how these rules were written). There’s three reasons, but one more than the others. There were more protected zones in 2020. The 2020 season was cut short, so we never really saw the full scope of what competitions may have brought. But most importantly, the game piece in 2020 was approximately the height of the top of the bumper zone, while 2022’s game piece was clearly taller than the top of the bumper zone. This meant that most of 2020’s intakes would extend slightly lower than bumpers, and thus would contact other team’s bumpers first when a collision occurred. But in 2022, most team’s intakes were fully above the height of bumpers, and thus would collide with opposing robot’s inside of their frame perimeter.

18 Likes
18

Minor nitpick - I think 2017 just had a max bounding box, but if your frame was smaller you could expand past the frame to the max bounding box - see 254 with an over-the-bumper intake.

4 Likes
19

R03. Maximum ROBOT size, including BUMPERS and all extensions, must be constrained to one of two volumes:
A. 36 in. by 40 in. by 24 in. tall (~91 cm by 101 cm by 60 cm tall).
B. 30 in. by 32 in. by 36 in. tall (~76 cm by 81 cm by 91 cm tall).
The ROBOT must remain constrained to the maximum inspected volume at all times during the MATCH (i.e. A ROBOT may not switch between volume A and volume B without being reinspected).

1 Like
20

Glad it wasn’t just me! I only went to offseason events for infinite recharge, so I wasn’t 100 percent sure that what I was trying to say was correct. I think another thing that makes this issue even worse is that a lot of teams put their most vital electronic components(RoboRIO, PDH) low on their robots within easy range of any intakes that may end up within their frame perimeter.

While this may be enough to make some teams move these components away from this area, it still doesn’t make the robot completely safe. A prime example of this is Qualification 58 at the Hueneme Port Regional, where 3128 ran into 1678 about 1 min and 35 sec in. Both robots were obviously trying to intake a cargo, and the intake of 3128 happened to rip out a CAM wire(I believe) on 1678, even though they have their RIO and PDH mounted underneath. Compared to some of the defense being played during this season, this collision was not anything egregious, but 3128 still walked away with a red card.

The only solutions to the problem this season would have been to increase the bumper height so that the majority of intakes would end up bouncing off of bumpers, or to reduce the diameter of the cargo so the same effect would take place. There should be no reason that 2 teams obviously both trying to intake a ball collide and one ends up with a red card in any situation.

1 Like