Re: Hot Goal Timing Issues
 

We ran into the same issues as our autonomous does roughly the same thing.

It was also hard to fix the problem because there was so much inconsistency in the lag/delay of the hot display.

Re: Hot Goal Timing Issues
 

In LA it was really inconsistent, sometimes it would switch right away, sometimes it took a second and a half, leading us to miss our hot goal shot if the goal we were aiming for was hot second.

As someone mentioned above, this could easily be solved if the indicators started the match in the "not hot" position and flipped down to indicate hot rather than the other way around.

Re: Hot Goal Timing Issues
 

Video : http://youtu.be/rMa97wE1XIc

Week 4 was our fist event, and having read about the issues and seeing it on the videos we worked hard to get this worked out on Thursday practice matches.

We are using a sensor to detect the dynamic reflector, we line up at the start of the match, and we have to make our hot detection prior to moving the bot, once we move the photoeye is no longer aligned.

We chose this as we have never done vision processing on a comp bot yet, and this seemed so simple that it was not worth the effort, for the point bonus.

This was taken at Week 4 Midland Mid Michigan event, and the timings below were calculated with raw video from my Nikon D5100. I am not sure about the frame rates after processed by YouTube so posting data here.

Summary, we have to wait a full second to detect, it was 750 mils to see the dynamic sensor move in frame. Also the second between 6 and 5 on the clock is significantly longer than a second. I suspect as the field electronics are running the code to change the states of the dynamic sensors and lights, there is extra processing happening. It appears that the added length from 6 to 5 is the taken off from 5 to 4 as 6 to 4 timing is 2 seconds.

For our bot, 1 second after autonomous starts, we check the sensor, and then turn on the underglow to match the goal we are shooting at.

Re: Hot Goal Timing Issues
 

When did the 2nd hot goal light? Hard to tell from Youtube.

Thanks for taking these measurements.

Re: Hot Goal Timing Issues
 

The field update for week 4 seems to have made all the hot goal delays less than one second. They've also changed the timer on the video display so that it turns yellow at 30 seconds.

Re: Hot Goal Timing Issues
 

Thanks for collecting this data.

I don't know whether I should laugh, or slam my head into my laptop.

Re: Hot Goal Timing Issues
 

Laugh first, head slam second and hope it knocks you unconscious so you don't have to worry about it anymore.

Re: Hot Goal Timing Issues
 

Frame 222 - Right Goal Only lit
Frame 223 - Both Goals lit
Frame 224 - Left Goal Only lit

I shared the raw video file on my goggle drive, you should be able to click the link to download. Not sure if google will try to play or not, hopefully you can download.


https://drive.google.com/file/d/0B9P...it?usp=sharing

*when you click the link, you can use the "File" menu to download.

Re: Hot Goal Timing Issues
 

Also, picture of the actuator for the dynamic reflector.
I would guess this reacts much quicker than the 750 mils. :)
Pretty sure it isn't hardware.

https://www.facebook.com/photo.php?f...type=3&theater

Re: Hot Goal Timing Issues
 

Team 3018 was in 18 matches week 4 at Wisconsin (19 autonomous periods since one match had a replay declared partway through telop).

We hit the hot zone in 18 out of 19 attempts. The reason for the one failure was that the hot zone flapper stayed hot on the non-hot goal for the first 5 seconds. We hit the shot in the first 5s because that is what the flapper told us to do.

The robot can score in about 1.8s from the white zone so the auton algorithm is to wait for 2 full seconds before checking for the hotzone (we use vision tracking of 4 pics per second). It begins moving as soon as it sees a hot marker or until a timer set at autonomous_init() hits 5s.

I've seen us start moving *before* the hot goal has moved to our side but it is always hot by the time we hit it. I think this must be due to the long 5-6s interval.

At our first event (Lake Superior week 2) we had unbelievable issues with the hot zone. Originally we waited only 1s and we were using the drive station object (java) to check match time to decide when to go. We were always going early so we first had it start waiting 2s instead of 1s. But we would still always go early. We set our own timer in autonomous_init and things started working like at home, but we kept the 2s wait 'just to be safe'.

Then we discovered the flappers at one end of the field were reversed, and later we had another stuck (both stay hot) scenario. We lost a match by 3 points in one match where the field lied to us about which was hot (we'd have been the first seed instead of second had that worked properly). Finally, (completely unrelated to to hot zone targeting) we had a physical catch to keep our robot in frame that a motor had to break away from. By elims we were bent and battered so even though the robot was moving for the hot zone, the breakaway failed for mechanical reasons (catch too firm all of a sudden) and the shot never happened. We only scored 5pts for about 6 matches in a row. A piece of scored electrical tape allowed us to function properly in the very final match.

Things were certainly better week 4 at Wisconsin (we used a magnet base breakaway which completely solved the one problem), but not quite perfect due to the one bad flapper event.

It's only 5 points, but it can be the difference in some matches, so it is nice when it works properly.

Re: Hot Goal Timing Issues
 

Those situations, if they occur again, should be brought up to your FTA - those sound like prime candidates for field faults.

Every morning the field staff should run a full test where they check for this sort of thing, so they should be pretty rare.

Re: Hot Goal Timing Issues
 

Seriously?

Coding timers that remain true to real-time while other stuff is going on is not difficult. Seems to me that the field timing should probably have 'remains true to within 100 ms throughout the match' as one of its mission critical requirements.

Its stuff like this that makes me seriously question both the people *doing* the work, and the people *checking* the work in Manchester.

Re: Hot Goal Timing Issues
 

We brought the issues up to the FTA at Lake Superior. The one failure at WI occurred in our 2nd quarter final match and it did not affect the outcome. My son is the lead programmer and was driver and he saw the problem. I'm not sure if he mentioned to the FTA or not, but we started in the same place for finals and it was fine.

At Lake Superior, we had to review video, check code, etc. before we found the root cause of the problems. The reversed flappers I think were only problematic on practice day. The stuck flapper required us to review the video before we found the problem and by then it was too late to ask for a fault. (I think we did ask and were told it was too late-- it was a couple matches later).

None of the problems hurt us from winning Lake Superior, nor reaching finals at Wisconsin so we're not grousing about it-- just sharing our own observations.

We now *always* watch the flappers (both left/right) every match as you never know when those 5 points will become critical and unless we notice it live it will be too late to go to the question box. Even then, I'm not sure it will result in a replay as it is unlikely the refs are watching for it and they won't accept video evidence.

Re: Hot Goal Timing Issues
 

Any news on if the initial hot goal delays have persisted through week 5? If so, is the delay now consistently <1sec? Just wondering for autonomous implementation. Thanks!

-Mike

Re: Hot Goal Timing Issues
 

Looking at the images we recorded from DC indicates that there was still a delay.