Non-level bumpers

[Chris is me]

"No" is unambiguous, but if bumpers have to be parallel to the floor, why does the manual say bumpers don't have to be parallel to the floor? Why not just change the blue text to say "bumpers must be as close to parallel to the floor as reasonable"?

Obviously angled bumpers are illegal if they say they're illegal, but I don't understand how the GDC would expect teams to interpret identical wording differently in different years.

[FrankJ]

Q199 said the bumpers had to be parallel to the ground. It also said the different bumper segments could be at different heights.

Maybe that gives enough room to creatively build the robot to meet your needs without lawering the rules?

[Alan Anderson]

I don't believe the GDC changed their mind between last year and this year. Nothing I've read indicates that the 2013 bumper rules permit angled bumpers. I think the Lead Robot Inspectors who gave angled bumpers a pass last year were not interpreting the rules the way they were intended. Since it was in teams' favor to accept that interpretation, nobody went up the chain of authority to get an ultimate ruling, and the LRI's word was final.

The answer from the GDC this year is clear: bumpers should not be angled. It's basically the same answer they gave last year, with the addition of the explicit "no".

[magnets]

Originally Posted by FrankJ

Q199 said the bumpers had to be parallel to the ground. It also said the different bumper segments could be at different heights.

Maybe that gives enough room to creatively build the robot to meet your needs without lawering the rules?

Our bumper sections are 0.5" long, aren't angled, and function as an angled bumper would for our intake.

Your dx approaching zero comment gave us the idea. Thanks!

[BBray_T1296]

We all do need to remember that the 2013 rulebook is completely nullified and 100% irrelevant for the 2014 game. Any rulings made in 2013 do not "roll over" into the new season. Just because they can Copy/Paste year to year, it doesn't mean we get to.

And next year in 2015, any arguments that are based on "how it was in 2014/2013/1992" will have no traction.

IMHO, having angled bumpers (relative to the vertical plane) is quite ridiculous. I would compare it to people in 2012 trying to "balance" on the bridge by driving up the Lexan guards underneath.

[Al Skierkiewicz]

Alan,
This specific ruling was discussed during multiple LRI meetings and conference calls. The answer in 2013 was any and all bumpers located between 2" and 10" above the floor were legal. I personally inspected several at each event.

[Abhishek R]

Originally Posted by magnets

Our bumper sections are 0.5" long, aren't angled, and function as an angled bumper would for our intake.

Your dx approaching zero comment gave us the idea. Thanks!

Bumpers designed using calculus, well played.

[GeeTwo]

Well, I'm both proud and ashamed to say it, but we finally picked up a ball tonight, and dropped it into fairly stable well where we could "kick" it over the truss or into a goal. (Proud that we did it, and ashamed at how little time is left to tweak and tune the controls and give the drivers practice) We did it with vertical plywood backing the bumpers, and horizontal bottom lines and pool noodles (at 2.5" off the floor to the lower edge) coming in 8" from each corner. We did do some interesting miter/bevel stuff with the track INSIDE the frame perimeter once we pulled it over the bumpers, and the final mod that made it work was made in time-saving/desperation, and involved using pool noodles inside the frame perimeter. The noodles were intended as a filler, but they apparently also served as a traction device, moving us from "almost working" to "that's it". Given the late date, we're locking this in as a solution, and are forging ahead towards stop build day.

[Dominick Ferone]

In the Q&A I asked
Q. As Per rule G27 robots can not use a wedge mechanism in order to flip other robots, but can we put our bumpers on an angle in order to help aid the ball in rolling into our chassis?
2014-01-31
A. BUMPERS must be backed by the FRAME PERIMETER and may not go inside the FRAME PERIMETER (see Figure 4-5). Please see the Blue Box on R2 for help determining the FRAME PERIMETER of the ROBOT.

and with the response it seems to sound like they dont approve of it but they dont say no to it but i could be wrong.

[Alan Anderson]

Originally Posted by Dominick Ferone

]...can we put our bumpers on an angle in order to help aid the ball in rolling into our chassis?
2014-01-31
A. BUMPERS must be backed by the FRAME PERIMETER and may not go inside the FRAME PERIMETER (see Figure 4-5). Please see the Blue Box on R2 for help determining the FRAME PERIMETER of the ROBOT.

and with the response it seems to sound like they dont approve of it but they dont say no to it but i could be wrong.

I believe you are wrong. The FRAME PERIMETER is not angled. A bumper backed by it will also not be angled. An angled bumper will not be backed by the FRAME PERIMETER and will not comply with the rules.

[GeeTwo]

To be more specific, there seems to be no question that bumpers angled from the vertical (that is, like cow catchers) are illegal. As for bumpers angled with respect to the horizontal (that is, across the frame perimeter): they were allowed in 2013, but they are questionable to forbidden in 2014 based on the Q&A, despite the rules being the same. Unfortunately, none of the questions has clearly defined the axis of slope, so at this point, the ruling is uncertain. Since I feel that my team's bumpers are beyond reproach, I'm not going to put this question to the judges; if you're not so sure, you should ask a VERY specific question on Q&A.

[Tristan Lall]

Originally Posted by Alan Anderson

I believe you are wrong. The FRAME PERIMETER is not angled. A bumper backed by it will also not be angled. An angled bumper will not be backed by the FRAME PERIMETER and will not comply with the rules.

There's nothing in the rules that requires that the frame perimeter polygon be parallel to the floor (if the robot is appropriately shaped). Also, it is uncertain whether a skew polygon (as opposed to a planar polygon) satisfies the rule.¹

Also, while the rules require that (certain parts of) the bumpers be backed by the frame perimeter, that doesn't create a constraint in which the entire vertical projection of the bumper has to lie directly above that locus of intersection. For example, one might conceive of a twisted bumper that was vertical at the ends and reclined in the middle. (But R21 and figure 4-8 may have something to say about that, on different grounds.)

Originally Posted by Dominick Ferone

In the Q&A I asked
Q. As Per rule G27 robots can not use a wedge mechanism in order to flip other robots, but can we put our bumpers on an angle in order to help aid the ball in rolling into our chassis?
2014-01-31
A. BUMPERS must be backed by the FRAME PERIMETER and may not go inside the FRAME PERIMETER (see Figure 4-5). Please see the Blue Box on R2 for help determining the FRAME PERIMETER of the ROBOT.

and with the response it seems to sound like they dont approve of it but they dont say no to it but i could be wrong.

Originally Posted by GeeTwo

To be more specific, there seems to be no question that bumpers angled from the vertical (that is, like cow catchers) are illegal.

That Q&A response doesn't follow from figure 4-5, but R21 and figure 4-8 may offer clearer guidance, if interpreted using the same standard as in R22. Of course, consistency of enforcement will be difficult, given that FIRST's guidance (about another, similar rule) is to look for overt deviation.

¹ Would FIRST really want to prevent a robot from passing inspection because its only possible frame perimeter is a skew polygon? Would it serve any purpose to interpret the rule that way? As usual, I'm inclined to give teams the widest possible latitude within the precise letter of the rules. And given that FIRST already allows curvilinear polygons, I see no particular reason to infer that they object to skew polygons.

[FrankJ]

As I said before I am not the one that is going to be inspect your robot. The person inspecting your robot might not have passed calculus. You robot inspector might also reasonably expect you bumper segments to be discrete. As Dx approaches 0 you might run afoul on the rule that requires you to be able to change your bumpers in a reasonable amount of time. Also keeping your numbers in order, (that would have to span extremely small bumper segments), will be problematic.

As for an angled frame perimeter goes. In order to be on the frame perimeter, when you push your robot against a wall, the point must be able to touch the wall and be in the bumper zone. If a point cannot touch the wall, it is not on the frame perimeter. (absent the minor projections of course.)

[Al Skierkiewicz]

Tristan,
I am confused by your response, the FRAME PERIMETER is a series of planes, that are essentially perpendicular to the floor when the robot is in the Starting Configuration. As such, the backing of the bumper system that is mounted on the FRAME PERIMETER must also be perpendicular to the floor. There is no occasion where any part of the bumper system can be skewed or angled with respect to the plane(s) of the FRAME PERIMETER. The FRAME PERIMETER is determined by the outer most section of the Frame as determined by string test in the bumper zone. Once the FRAME PERIMETER has been established, it can be measured and the bumper system can be inspected. It is not possible to meet the rules with respect to bumper mounting (i.e. backed by frame with no gap greater than 1/4" be wider than 8" and supported at the ends by at least 1/2") on a skew polygon unless said polygon was skewed vertically by 1/4" or less in the bumper zone.

[Tristan Lall]

Originally Posted by Al Skierkiewicz

Tristan,
I am confused by your response, the FRAME PERIMETER is a series of planes, that are essentially perpendicular to the floor when the robot is in the Starting Configuration. As such, the backing of the bumper system that is mounted on the FRAME PERIMETER must also be perpendicular to the floor. There is no occasion where any part of the bumper system can be skewed or angled with respect to the plane(s) of the FRAME PERIMETER. The FRAME PERIMETER is determined by the outer most section of the Frame as determined by string test in the bumper zone. Once the FRAME PERIMETER has been established, it can be measured and the bumper system can be inspected. It is not possible to meet the rules with respect to bumper mounting (i.e. backed by frame with no gap greater than 1/4" be wider than 8" and supported at the ends by at least 1/2") on a skew polygon unless said polygon was skewed vertically by 1/4" or less in the bumper zone.

The frame perimeter definition refers specifically to a polygon, not to a series of planes:

Originally Posted by Section 6.1

FRAME PERIMETER: the polygon defined by the outer-most set of exterior vertices on the ROBOT (without the BUMPERS attached) that are within the BUMPER ZONE. To determine the FRAME PERIMETER, wrap a piece of string around the ROBOT at the level of the BUMPER ZONE - the string describes this polygon.

Note: to permit a simplified definition of the FRAME PERIMETER and encourage a tight, robust connection between the BUMPERS and the FRAME PERIMETER, minor protrusions such as bolt heads, fastener ends, rivets, etc. are excluded from the determination of the FRAME PERIMETER.

Perhaps the series of planes to which you refer is the "vertical projection of the FRAME PERIMETER" (although that could include curved segments which are not strictly planar):

Originally Posted by Section 6.1

STARTING CONFIGURATION: The physical configuration and orientation of the ROBOT when the MATCH is started. This is the state of the ROBOT immediately before being Enabled by the Field Management System, before the ROBOT takes any actions, deploys any MECHANISMS, or moves away from the starting location. This configuration is static, and does not change during a single MATCH (although it may change from MATCH to MATCH). In the STARTING CONFIGURATION, no part of the ROBOT may extend outside the vertical projection of the FRAME PERIMETER, with the exception of minor protrusions such as bolt heads, fastener ends, rivets, etc.

If a ROBOT is designed as intended and pushed up against a vertical wall (in STARTING CONFIGURATION and with BUMPERS removed), only the FRAME PERIMETER (or minor protrusions) will be in contact with the wall.

Since the bumper rules refer to the frame perimeter, but not to its vertical projection, I have to assume that they should be evaluated with reference to that polygon (whether it be rectilinear or curvilinear, and planar or skew¹).


To expand on what I'm talking about, consider a robot with four sides, each with an outermost edge running the whole length of the side within the bumper zone (i.e. the major protrusions relevant to the frame perimeter). If in the starting & playing configurations, the left and right edges are at 9.5 in from the floor, and the front and back edges are 2.5 in from the floor, there's no planar polygon that can satisfy the string test. I'm skeptical that FIRST means for that robot to be illegal, and the only way to maintain any measure of fidelity to both the rules and geometry is to allow a skew polygon—i.e. a figure composed of segments that do not necessarily lie in a plane.

That's the "skew" to which I refer, but when I speak of "twist" I mean something else. When I describe a twisted bumper, I conceive of one in which the backing material is bent out of the vertical plane. I argue that if the bottom edge of the bumper backing lies on the frame perimeter polygon, and the top edge lies inside of it, then R21 is satisfied, except—depending on interpretation—for the part about the vertical cross-section (including figure 4-8).

Now, as for the vertical cross-section, R22's blue box offers a suggestion about how to interpret figure 4-8 in a different context:

Originally Posted by R22

BUMPERS must be located entirely within the BUMPER ZONE, which is between two (2) and ten (10) in. from the floor, in reference to the ROBOT standing normally on a flat floor.

There is no explicit requirement that BUMPERS be perfectly parallel to the floor, however the requirement that BUMPERS be constructed per Figure 4-8, the vertical cross-section, does implicitly mean that a BUMPER should not overtly deviate from this orientation.

I think it's fair to apply the reasoning found there to R21, even though the blue box is attached to R22. After all, they're referencing the same figure and FIRST has never given any indication that they mean for a distinction to exist. The Q&A seems to be consistent with that interpretation of the constraints imposed by R21.

But note the distinction between "may not go inside the FRAME PERIMETER (see Figure 4-5)" and "should not overtly deviate from [vertical]". Figure 4-5 does not establish any constraint upon verticality (or flatness, for that matter). If they meant instead to refer to figure 4-8, and we interpret it as above, we're still left with a question of how much deviation is overt. That's why I express concern about consistency of enforcement: if it's 10° at one event, and then 2° at another event, we might have a bit of a problem.

Finally, "may not go inside" is phrased very strongly (compare "should not overtly deviate"), and this is an impossible constraint if there's any twist at all—as one might expect in a real physical part. I therefore contend that the Q&A is probably wrong in that respect, because we can't reasonably presume an intent to impose a practically impossible constraint upon a mandatory feature. Accordingly, I fall back on the necessity for officials to judge overt deviation.

¹ To avoid skew polygons (and other issues), I've previously suggested that the frame perimeter definition could be improved if it were based upon the convex figure formed by the projection of the outermost extents of the robot (within the bumper zone) on to the floor, in the starting configuration.