Minibot's triggering targets?
 

As I was watching a few matches today I noticed many minibots being awarded wins while the lights on the towers remained racing and not "triggered" to indicate finishing place.

<G02> The AUTONOMOUS PERIOD ends when the ARENA timer displays zero seconds. The MATCH ends if all TOWERS are TRIGGERED or when the ARENA timer displays zero seconds, whichever comes first.

During the AUTONOMOUS PERIOD, the BASES will be illuminated in yellow. During the TELEOPERATED PERIOD, BASES will be illuminated with their ALLiANCE color. At 15 seconds before the end of the MATCH, BASES will flash yellow, while TARGET lights will flash green. At 10 seconds before the end of the MATCH, when HOSTBOTS may DEPLOY their MINIBOT without penalty, BASES will illuminate with the appropriate ALLIANCE color, while TARGET lights will show a green „chase‟ sequence. This chase sequence will continue until a MINIBOT TRIGGERS the TARGET, or time expires, whichever comes first. If a MINIBOT TRIGGERS the TARGET within the set time period, the TARGET light will illuminate to indicate in which place the MINIBOT finished. 1st place will illuminate all four lights to indicate the maximum number of points have been scored. MINIBOTS finishing in subsequent positions will illuminate fewer lights, with the 4th place MINIBOT illuminating one light, to indicate the fewest points scored

And Triggered from Section 1 is defined as:


TRIGGERED – the act of pushing the bottom disk of the TARGET so that the sensors are tripped and a signal is sent to the Field Management System (FMS). When a TARGET is TRIGGERED, the MINIBOT RACE on that TOWER is complete.

Was the system not working so the places needed to be done visually or are the minibots designed wrong so that they don't "trigger" the targets as required?

What did you see at your regional?

Re: Minibot's triggering targets?
 

This is second-hand information, but some of the BAE volunteers were saying that the towers sometimes trigger when a robot hits them hard enough. They're working on a solution, but to avoid false triggers they've been doing it manually, which means the lights don't always work.

I don't know if this is the same reason at the other regionals.

Re: Minibot's triggering targets?
 

Nutty,

This is what I wanted to know. Have no problem with it if the case is the triggers are not quite right yet. Hope it gets fixed before we have to call a really close race with 4 minibots at say <1 sec...

Re: Minibot's triggering targets?
 

Mini bots were not triggering the towers at the Alamo either. There are plenty of refs to watch for the towers being triggered but I have no idea how they would call a close race.

Re: Minibot's triggering targets?
 

Only a couple minibots triggered the towers here at traverse city so far.

Re: Minibot's triggering targets?
 

So of course as always, this is completely "unofficial".... but:

There's a "feature" in the field PLC network to tune the triggering on the towers to differentiate between base bot slams and the minibot top hat triggering. Needless to say that its a tough one to crack as when you think about the bots hitting the base hard, and then you have a 10' lever to throw around that switch plate at the top. So FIRST FRC Engineering Ryan at the NJ regional has been playing around with the tuning all weekend trying to figure out what works best as testing this in the lab was next to impossible being we didn't know how the competition bots would be interacting with the fields. Seems like he's finding the happy place for it. As always, refs have final authority on scoring...

Just like you guys know, no real substitute for Regional competitions, we've seen a few things pop up this week that didn't happen in Week 0 events, it'll all be sorted for Week 2, gotta love engineering development, keeps you on your toes...

Re: Minibot's triggering targets?
 

At FLR they were doing it manually. There were some very close minibot races, and I have no clue how the refs called them.

Re: Minibot's triggering targets?
 

Also, a lot of minbots (like ours) sometimes just don't have enough force to trigger them because the switch will turn off the minibot right when it hits the plate, so it can't go a bit further to push the actual triggering mechanism

Re: Minibot's triggering targets?
 

Then your minibot didn't really meet the challenge, correct? From section 2.2.5:

A minimum contact force of approximately 2-4 Newtons, depending on contact location, is required to ensure the contact sensors in the TARGET trip reliably.

Re: Minibot's triggering targets?
 

That's what I'm wondering as well. My completely ancedotal and likely biased evidence is that it was mostly lightweight bots at Alamo that were failing to trigger towers. I can only trust that correct calls were being made per the rules as written, but I'm doubting refs can or are capable of visually judging the force exerted by the minibot on the plate.

What I am certain of is that the current lack of information can only lead to confusion and hurt feelings. When a team loses an important match due to a "phantom" minibot deployment, they're going to be rather annoyed. If someone from the gdc or elsewhere could clear this up it'd save a lot of hurt feelings and comparisons to the infamous "tape measure" rule.

Re: Minibot's triggering targets?
 

Yeah, I've got similar concerns. While I of course appreciate the difficulty of making the mechanism trip reliably, the rule is written such that it's not the contact of the minibot upon the pressure plate, nor the force, but rather the act of tripping the sensor that completes the act of triggering (and ends that minibot's race).

If the refs are calling mere contact, they're doing it wrongly. And yet, it's impossible for them to do it right if the sensors aren't working—and nearly impossible to do the next best thing, which is render an accurate estimate of the sensors' states.

Referees can estimate the force (based upon the estimate published by FIRST), or guess at the deflection (approximately 0.25 in is necessary), or use some other proxy for triggering—but according to the rules, all that matters is that "the sensors are tripped and a signal is sent to the Field Management System". If the sensor doesn't trip because of some aspect of the design of the tower, then the alternatives are field fault (if it should have tripped), or no call (so that may be a feature, not a bug). For a given minibot design, how do you know which is the correct outcome? And on top of that, there are four towers, and the order matters! How do you judge all four at once?

I don't blame the refs—this one needs a rule change to grant them the liberty to make a different call (i.e. to allow them to call it based on contact alone).*

And of course, if a hostbot causes the sensor to trip by contact with the base, it isn't triggered, and in an ideal circumstance, the field should still allow the minibot race to take place on that tower, and allow the sensor to be tripped properly. If a minibot is prevented from concluding its race...I would have to call that a field fault too. (An insufficient degree of fault tolerance changed the final score of the match.)

*Can they bend the rules in the best interests of the competition? Maybe just a bit...because their objectives include some sort of equity. But that will end well only if FIRST is on board. Referees (more than inspectors) have to really be careful about this, because of the increased exposure and shorter timeframe available for deliberation.

Re: Minibot's triggering targets?
 

Along the same lines, in talking to one of the technical volunteers at FLR, it sounds like there is a minimum dwell time for the top plate (I assume to implement the differentiation between base slams and minibot touches) - the fast minibots may not be deflecting the plate long enough for the current field system to recognize it (BTW, there isn't anything in the rules (thankfully!) for how long plate would need to be deflected).

The complication is that end of match can also be triggered by all four towers being tripped, so false positives can be REALLY BAD (tm).

At least at FLR, the response was to use lots of eyes - the refs, the head ref, and addtional volunteers. AFAIK, they were watching plate motion (which is what the rules actually call for), and not just order of finish. Confusing for the for audience (I hadn't heard any announced explantion, but there may have been one.), but the officials seemed to do well.

Sounds like the engineering team is on top of this, and hopefully will have it fixed for next week.

--Karlis

Re: Minibot's triggering targets?
 

According to this news report, the Finals at BAE GSR were actually contested due to this exact issue:

http://unionleader.com/article.aspx?...d-a4aac9a692ba

We were one of the the teams at BAE with a minibot that always climbed but never triggered the tower properly. We use a common Decora-style light switch on the top with hard physical stops on either side of it to protect the switch from being "crushed" into the plate when it hits the top. Our minibot definitely hits with the required force, but, as mentioned above, we suspect it may be the dwell time at the top that is the issue. The head of robot inspection came by our pits to specifically check our triggering setup on the minibot, and he said they were doing a survey of different triggering systems in an attempt to determine what did and did not work properly.

In any case, I would hope they figure this out quickly and fix it. The article above quotes an official as saying "We don't do video review here", but maybe that's something FIRST should actually consider.

Re: Minibot's triggering targets?
 

TARGET design fail.

Solution. Should be using either counter sunk heads on the target bolts or switch to a shoulder bolt. The issue i saw was minibots coming into contact with the pan head bolts, the hitting at this off angle it seemed to be getting caught on the thread that make it move up.

Our 5#'s of slamming minibot into it would not trigger it. At Kettering many minibot races were called by eye.

Re: Minibot's triggering targets?
 

They came around and showed us at kettering that there are bolts that stick down out of the trip plate that don't move.

So if your robots hit those bolts the plate will not move and the tower will not be tripped.

The plates were not designed right. The bolts should have been mounted solid to the bottom plate and slide up through the top plate so that no mater where you hit the plate would still be tripped.

At kettering they used stop watches at each tower and confirmed against the computer. They field crew was very busy all weekend and should of had more help. Between the field comunication issues, driver station issues, and minibot issues. There was no time for them to help teams with communication issues with the field. Also there where only 2 practice rounds the whole event, meaning only about 12 of the 40 teams got to make one practice round. Everyone else didn't get a single practice round.

Re: Minibot's triggering targets?
 

After having spent many hours acting as a high school swimming timer, I can tell you that this is a similar situation. There are "touch" pads at the end of the lane to automatically time the touch, there are people standing at the end of each lane with a manual trigger that is pressed when the swimmer touches, and there are two people with stopwatches that back-up the automatic system. Simple triple redundancy. If everything works, the touch-pads are the official score. The manual trigger is the backup for the touch-pads. In the event of a total system failure, the stopwatches are the official time.

Why not do something like that?

Re: Minibot's triggering targets?
 

For those of you insinuating small bots are not triggering, that was definitely not the case at FLR. The switches at the top of the pole are simple limit switches that deflect easily. The false positives during practice at FLR caused the field team to basically disable (I don't know what they actually did, but only one pole was really triggering) the tower finish lines. The mini bots traveling up the pole in 1.5 seconds or less have at least 10 times the energy required to trigger at the specified trigger force.

I am sure the system will get fixed, but the tower is definitely the problem right now.

Re: Minibot's triggering targets?
 

We had a lot of issues triggering the poles at Kettering. Fortunately the Refs were reasonable about this, and there were only a few actual minibot races that there was a concern of which alliance got what place. The switch on our bot had a push force at 1.5X the maximum force specified in the manual. The bot itself builds up enough momentum to make the 10 foot steel pole on the practice field jump (not properly anchored) when it strikes. Force is definitely not the issue...

Re: Minibot's triggering targets?
 

The fault very much lies in the sensing equipment. I saw far more triggered by robots colliding with the the pole rather then minibots hitting the sensor.

Re: Minibot's triggering targets?
 

Indeed, the sensors seemed to be working fine in practice, but appeared disabled for qualification rounds. They were fooling with them often during the days, but the manual system worked well.
I don't think so. The minibots were easily hitting hard enough to trigger the switches, more than hard enough to overcome incidental friction from screw threads. My suspicion is the voting system "wasn't quite right". By the end some towers indicated consistently, but that left rear one never, ever did.

Re: Minibot's triggering targets?
 

I asked a ref about it at Traverse City, and he cited the "dwell time"; the programming for the field makes it so that the sensor plate doesn't trigger unless it is held for a short time (long enough that hitting the base won't trigger it). Ours triggered it because it held the sensor for a short time.

Re: Minibot's triggering targets?
 

Interesting design choice. If the FMS is handling that debounce routine, those are field faults: the sensor tripped and sent a signal (hence the tower has been triggered), and the FMS ignored it, because of an arbitrary timer with no basis in the rules. If, instead, the debounce is being handled on the sensor (so the FMS doesn't get a signal), then the fault lies in not accurately describing to everyone how the sensor actually works.

Ultimately, I think we've got to see a rule change and/or a design change, because it looks like the field is implemented in a manner that encourages field faults, and the only solutions available to the referees are by definition complete subjectivity with a large margin of error.