TEST Einstein BEFORE the BIG-show... here's a proposed approach...

[techhelpbb]

Originally Posted by qzrrbz

I don't think this necessarily follows. This doesn't rule out either environmental dependencies or data dependencies.

Take the assertion that the field was tested as a given, but tested when and how? If at o'dark30, then the sea of wifi enabled phones in the audience weren't there. The testing environment is not nominally equivalent to that seen during finals, and I have no idea how it could possibly be made so.

The data requirements were probably not as seen in finals, but this, I'll grant, is "in the robots" -- but it's possibly *not* their fault.

This is a hard problem, no doubt about it. Intermittent and barely reproducible -- a troubleshooter's worst nightmare!

To some extent the spread spectrum nature of the radios and the fact that many phones are not 5GHz should have eliminated a lot of that.

This is not to say that there can't be an impact from 2.4GHz traffic, it's not entirely clear to me that my definition of disabled equals what FIRST considers disabled when they discuss the 2.4GHz portion of the D-Link AP radio.

Then again, I appreciate that perhaps the robot traffic simply was too much, under the cirumstances I can see that happening.

[DjScribbles]

I had a similar thought while the field was having issues:

I think the field should include 6 (maybe less for cost reasons I guess) CRIO's hooked up to 6 radios, configured as imaginary team numbers 9996-9999. Program each CRIO in a different language and fit it with some basic code (or maybe exhaustive code using alot of functions as a 'test') and these boxes become the gold standard for proper field functionality.

If a field has issues, hookup the boxBots, if they don't work then you know it's the field; if they do work then that gives the field personel a little more certainty that the field is not at fault.

It's a bit of an expensive solution to start, but these can be used multiple years with fairly minimal maintenance, and it removes a variable from the equation, because it's easy to blame the robot/team that had problems.

(edit for sounding less like a jerk having a rookie bracket of sorts is a cool idea; but I don't know that it would be able to solve field problems in advance, I suspect having a rookie team fail on the field is more than likely going to be attributed to that team than the field).

[techhelpbb]

Originally Posted by DjScribbles

Here's what I don't quite get... you don't need a robot or a bunch of teams to test it; that just opens the door to brushing off any problems as being the fault of the participant.

I think the field should include 6 (maybe less for cost reasons I guess) CRIO's hooked up to 6 radios, configured as imaginary team numbers 9996-9999. Program each CRIO in a different language and fit it with some basic code (or maybe exhaustive code using alot of functions as a 'test') and these boxes become the gold standard.

If a field has issues, hookup the boxBots, if they don't work then you know it's the field; if they do work then that gives the field personel a little more certainty that the field is not at fault.

It's a bit of an expensive solution to start, but these can be used multiple years with fairly minimal maintenance, and it removes a variable from the equation, because it's easy to blame the robot/team that had problems.

There's a bunch of issues with testing radio issues at these frequencies.

The orientation of the robots matters.
The 3D locations of the robots matters.
The sort of traffic you are trying to push through matters.
The internal designs of the robots and radio placements matter.
What the radios can and cannot do, like disable 2.4GHz fully matters.
Even the projectiles on the field, and the people near it matter.

The trick would be to basically create good approximations of the robots.
The difficulty is knowing what those approximations should look like.

Otherwise the mock robots pigeon hole you into making your design more like theirs or you are back to the bad design example. Worse it's expensive.

At this point you may as well drive the robots you have with the teams that brought them. They gain experience and you can see if there's a little gotcha coming that a straight simple field test didn't show.

[JohnSchneider]

Originally Posted by akoscielski3

I would like to make a Huge point about the Einstein problems that people are assuming never happened. Here is is... Everyone listening!?!

THE FIELD WAS TESTED BEFORE EVERY ROBOT GOT ON THE FIELD!

I was told by a person who I know is extremely reliable about the Field issues, however I will not release their name to anyone.

/END ALL EINSTEIN ISSUE DISCUSSIONS!

I was told by a person who I know is extremely reliable and whose name I will not release that you made that up

[bduddy]

Originally Posted by akoscielski3

I would like to make a Huge point about the Einstein problems that people are assuming never happened. Here is is... Everyone listening!?!

THE FIELD WAS TESTED BEFORE EVERY ROBOT GOT ON THE FIELD!

I was told by a person who I know is extremely reliable about the Field issues, however I will not release their name to anyone.

/END ALL EINSTEIN ISSUE DISCUSSIONS!

Tested how? With 6 robots running around sending and receiving data? I doubt it...

[Dale]

I love the idea of somehow opening Einstein to the top seeded teams, rookie or otherwise, who weren't picked in their division. Maybe it's just the robots who would otherwise be on standby. Maybe leave the top ranked unpicked robot ready to roll at their home field and let the others play on Einstein. If the left behind robot is picked then you pull out the next one from the Einstein matches.

I know from experience that about the worst thing you can be is the 4th highest seed unpicked robot. You can't pack up but the chances of you playing are next to nil. This would give those teams a fun consolation prize.

[sgreco]

Originally Posted by nikeairmancurry

I don't see where you can get so much wear and tear after playing a few matches..

I agree, but if they don't want any wear on the field at all, they can test the field without playing matches.

In other words, what they really need to test is electronics and communications, and that can be tested without having physical robots beating up the field (whether that wear and tear is significant or not is up for debate, but if they wanted people to use the fields before the finals, they wouldn't make a whole new field just for Einstein, so they'd probably prefer if nobody used it before the finals.)

[M_Bergman]

Originally Posted by akoscielski3

I would like to make a Huge point about the Einstein problems that people are assuming never happened. Here is is... Everyone listening!?!

THE FIELD WAS TESTED BEFORE EVERY ROBOT GOT ON THE FIELD!

I was told by a person who I know is extremely reliable about the Field issues, however I will not release their name to anyone.

/END ALL EINSTEIN ISSUE DISCUSSIONS!

To mis-quote the internet: "Names or it didn't happen."

On another note, I feel that the idea of a JV-team finals on Einstein sounds great - even if it doesn't particularly test the field, it gives these teams a great experience, especially if it's for the top rookie teams.

[Johnny_5]

Just saw this in a post earlier but I wanted to add my understanding about the actual FMS software and the FMS Light we can download. FMS Light is a mess and it is very difficult to understand and trudge through.

Thankfully FRC engineering understood this from last year and modified a lot of the process. Now the scorekeeper or field power controller is the one in charge of starting the pre start to engage the AP to generate the next set of SSID's. And the head ref is the one that actually publishes the scores up on the big screen after the match is over. This is part of the movement over toward real time scoring.

So this year the direct user interaction with the FMS software was significantly reduced. From what I understand the main interaction was to fix any scoring mistakes or penalty issues.

So mainly the interface was not directly through the server running FMS but more the indirect action through the field PLC that controls the real time scoring and AB PanelViews for the refs

[Banderoonies]

regardless of testing. I think the Rookie-stein idea is awesome!

[ratdude747]

Here is an idea from another thread:

Give each Einstein alliance a "practice period" where they have say 3:00 of field time to themselves. The time can be used to confirm field functionality, calibrate sensors, practice strategies, and impress the crowd.

[bduddy]

Originally Posted by ratdude747

Here is an idea from another thread:

Give each Einstein alliance a "practice period" where they have say 3:00 of field time to themselves. The time can be used to confirm field functionality, calibrate sensors, practice strategies, and impress the crowd.

That's not a bad idea, but certain issues might only become apparent with 6 robots on the field.

[Johnny_5]

FIRST seems to really be going to the show atmosphere during the finals, so I doubt we will see an integration of actual match time be it practice or not before the actual Einstein matches.

The show mentality is a great way to handle it and it looks much more professional (when everything works) that way. But the FIRST needs to change the way things are done that everyone who operates the field and everyone who uses the field has 100% in both ways and the equipment that connect the two.