2022 and 2023 both had horizontal extension limits that were very important to core gameplay, and both games had many robots that broke these limits. I’m obviously not going to call anyone out, and I don’t believe any team is intentionally violating this rule for an advantage either. The only thing I want to highlight is how easy it is to accidentally violate the extension rule.
During inspection, RIs will check if a team’s robot can overextend. If it can, the team can just say that software prevents it from violating any rules, and that’s enough. I don’t believe RIs ever ask to verify this. If the robot overextends on the field, the violation is very harsh. This year it was a tech foul if while scoring, and 2022 overextending while climbing made the robot ineligible for hanger points. In practice though, these were never called unless the robot was seriously and noticeably broken. Overextending by a few inches isn’t noticeable, especially if the movement is part of a sequence, like scoring a game piece or climbing almost always is.
From what I’ve seen, the most common way these accidental overextensions happen is when simple software limits are implemented to prevent them, but transitional states are forgotten. Sometimes this oversight doesn’t actually matter and could be easily fixed, but other times the transitional state is what enables functionality.
Is this even an issue? As a senior and software main I admit I’m not very experienced and it could be a lot less common than I think. Even if it is common, a lot of the time the overextension is inconsequential anyways. I think my main annoyance with this is that teams can easily break this rule without ever knowing, and it is impossible to get called on. The violations for the “don’t overextend” rules are purely cosmetic. I don’t even think this makes most games unfair, but it has the potential to.
To fix this, demonstrating scoring/climbing sequences could be part of inspection. There are some issues with this, but I think it would go a long way to make sure teams are aware that they’re violating a rule.
As I recall we had to demonstrate our software limits during inspection. Of course there really isn’t a way to know if the set points you show in inspection are the same as the ones you use during matches, but there’s really never going to be a way around that besides teams being ethical
In our experience this year, in neither of our two events did the inspectors ask us show that we had software limits preventing an extension beyond the maximum during inspections, despite the fact that the robot was clearly physically capable of extending too far.
Interestingly enough, in one of the tournaments, half way through our practice matches, a robot inspector approached us in the pit and told us that “there are concerns that our arm reaches out too high and far”. The source of the “concerns” seemed to be left intentionally vague. He then asked us to demonstrate that our normal scoring routine was within the rules by executing it adjacent to a wall where he had taped off the legal boundaries. As it turned out, we were actually very close to the maximum as our arm deployed, so we tweaked the angle to eliminate any doubt.
This is exactly the kind of thing that I feel should be done during initial inspection. The key here is that you’re demonstrating the scoring routine, not just a single position.
I’m extremely surprised that there were events where this wasn’t tested. We were asked to show our software limiting at all 4 events we attended. It is actually true however that they never made us test our translational states for scoring. They also did it at a different time at every event. At our first event they put us 48 inches away from a fence next to the inspection station and told us to demonstrate the limit. At our second district event it was done in our pit. At district champs, they had yet another station for extension limit testing but told us not to use it and instead did it in our pit. They also did it in our pit at champs iirc.
Software limit tests were performed (where needed) at every event I was at this season. Generally speaking, I dislike using a wall for this - the last thing I want to do is tell a team to operate their robot in such a way that there’s a chance it could break - and driving your arm into a wall presents that chance.
I’ll typically lay a tape measure on the floor, stand off to the side with a plumb bob at the correct distance, and eyeball it, only looking more closely if I feel it’s extending too far (aka moving the plumb bob closer, after I’ve seen and understood the motion of the machine and the speed it moves so I can do so safely). It shows teams that we take this seriously, while not being strict to the point of risking damage to the robot.
I also had teams demonstrate their capabilities and limits at events I was an RI for this year. I only asked the ones that were right near the limits to demonstrate though.
There was one incident where referees asked an inspector to check a robot for overextension. The inspector did, and found out that yes, they were extending too far, at least by several inches. Their “hard stop” had broken, so it was no longer preventing the extension. They fixed it.
We had ours mechanically limited so we wouldn’t have any issues. Last year we were software limited for our first event which we weren’t 100% confident in so between events we implemented a mechanical limit
While I agree that inspection should be testing the software limits on extensions, I must say that I do remember over-extension being penalized even when not the result of broken robots, at least in 2022. I can think of at least four times that was done during climbing here in NC district and a couple of other times it was done during match play. It was actually pretty easy for the refs to catch it if they were paying attention at all (and ours always were, they’re good refs) since there was a tape line on the hanger frame and the same on the goal uprights specifically to mark maximum extension. They got called especially on the climb because that was a violation that could gain the violating team an advantage and the refs were looking out for it. This year was a bit trickier for the refs, to be sure, since they didn’t have much in the way of good gauges on the field to judge over-extension against except the grid or obvious and egregious violations (which admittedly did usually involve broken robots.) But, of course, the over-extension away from the grid wasn’t much of a problem nor did it affect the game generally, so this was also less of a problem for the opponents of those who may have violated it in minor ways. So I don’t think that this is really a rule that’s all that easy to violate accidentally unless you’re just being careless.
We had to demonstrate our software limiting in order to pass inspection. Anyone who was software limited and didn’t have to demonstrate it, well, should have had to.
In terms of cheating by using different set points, there’s a whole lot of points where we depend on teams to act ethically, and are by and large unable to catch them cheating unless something goes wrong with their cheating: using an illegal extra compressor to charge their air in the pits, adding ballast after being inspected for weight that puts them over the weight limit, illegal cheesecake, illegal wireless communication between drive coach and human players, convincing a team to throw a match, being convinced by a team to throw a match, using illegal batteries modded to look like legal batteries, bribing the refs… I’m not at all saying that teams are doing the vast majority of this, and I presume that the vast majority of teams are doing none of this, and chances are good that no teams have ever done some of this–but we have to take all that on faith. And given that, using the honor system on the whole “is the pit software the same as the field software” bit, unless it obviously isn’t, seems quite reasonable.
The robot passing the check at inspection is no guarantee that it continues to pass. Depending on the design, there could be sequences of motion where the limits are exceeded so the testing would have to be “very comprehensive”, requiring a lot of time. Teams are free to make changes in their software and these changes can affect the limits, intentionally or unintentionally.
I have seen 3 different experiences this year. At our first regional our extension was never checked and we were never even asked if it was software limited. At our second regional it was checked and we were asked if it was software limited, but the limit or extension itself was never checked. Also at our second regional a team close to us was forced to change a software limit because the RI thought it was too far extended. Kind of just depends on who inspects your robot.
My team’s software limits for extensions were never checked at any of our official events this year, from our two district events to DCMP and Champs.
That being said, I’m not sure this should be a part of inspection. Wouldn’t it be a pretty time consuming step? At our events, the extension checks were done outside of our pit inspection, at the robot inspector station during weigh-in. We usually don’t bring a battery, driver station, or tether to this weigh-in, and we may or may not have reconfigured our radio at this point. My question for teams that actually had inspectors check for software limits on extensions is: when was this done, and how much of a hassle/bottleneck was it? Was it done during weigh-in or later after the power-on check in the pit?
Our first competition we had to show the robot actually going to our farthest reach arm position. We actually had to show all possible positions of our arm and had one position where it slightly extended out both sides of the robot, which we adjusted the software to eliminate that issue.
The second competition had us just manually move the arm to the position we thought was where it extended to at its max, which seemed a bit sketchy since there was no way for us to accurately get all three joints to their exact angles.
For us it was done in the pit. I do recall that last year when they checked extension it was just at weigh in (presumably since most robots mechanically stopped overextension).
This is interesting because my team (3142) attended two of the same events as you (DCMP and Champs) but we were never required to demonstrate any software limits for extension.
I have always done these sorts of checks as part of the power-on check. Usually in the pit, although there have been times it’s been handy to do it separately at the inspection station (for example, there have been years where the rule isn’t a straight extension like this year, but rather “Robot may not extend beyond a 54in diameter vertical cylinder”… practically impossible to judge without having a circle marked out on the floor).
Pretty similar as far as inspection being slightly unwieldy with a few actuations to manage to show limits, and not having any specific test other than manually move to furthest 48" extension and measure with tape.
We ended up not testing vertical limits but these were outside the scope of play for us and we’d pay a bigger price being higher CG than anything. Before design changes we maybe could have used the upper arm position with a transitional limit but that didn’t get used later.
We didn’t have a software or hard limit vertically mostly because the arm joint tracking was prone to a chain skip and the absolute encoder didn’t get implemented. And we didn’t up using the over the top transition where it could violate vertical limits when multiple parts were extended, so almost never would it be a issue during the matches. We did have one single match where the gripper wasn’t working and drivers accidentally violated height to try to hold things easier, but it was also the match the robot tipped over which reiterated that fully up and extended was bad.
Last, height wasn’t specified to test limits in the inspection checklist, which I generally except as a refs decision with some potential re-inspection for repeated violations. I recall the 2022 height limits being similar in they were more on refs judgement during climbs.