Yeah, I completely agree the rule will likely get clarified/updated in the upcoming Team Updates, mainly because of the nature of the rule itself, it feels like it’s just something they forgot to add. I just wanted to mention that it is technically legal as it is now.
3 Likes
Fantastic analysis from my favorite team I’m not on.
Keep up the good work can’t wait to see what else you all come up with!
7 Likes
9 Likes
Well, luckily we expected the rule to change and didn’t design any prototypes for it. Will remove the skill from the list 
12 Likes
Hardware & Software Update
Note Following
Continuing from Sunday’s post, we have tried some more things with NOTE detection and alignment. For this, we have currently implemented the following logic:
- When detecting multiple notes, we first select the middlemost note.
- Then, it rotates until the selected note is in the middle of the screen using a P-controller on the x-coordinate of the middle of the detected object.
Unfortunately, the simple color filter is easily distracted by any other orange thing or red bumpers. We might improve this later in the season using machine learning models, but this is currently not worth the resources.
Auto Paths
We conducted two back and forth straight line tests with our 2023 robot TR-X. Meaning these tests are done using regular NEOs instead of Vortexes. These tests gave us the following measurements:
- 8 meters (~26 ft) in about 7 seconds
- 6 meters (~20 ft) in about 5.5 seconds.
Afterwards, we tried to test the feasibility of a 5 note auto. Starting with scoring the preload, then score the three notes in front of the stage, and finally picking up a note at the centerline. We will send videos of this 5 note auto soon…
Shooter Prototype Concept
We came up with the idea to see if a deflection plate at the end of the shooter will work à la frc118 in 2013.
We think it could be very easy to make with a couple of hinges. We think this would be very much worth testing. Setting the shooter at a fixed angle and then just changing the deflection flap to see if we can make other shots that way.
For those who don’t know how this shooter works, here is a quick video of 118’s 2013 robot:
Intake prototype
We’ve successfully assembled the intake prototype! Coming up next, we’re planning to connect the CAN tomorrow. Once that’s done, we’ll move onto checking the programming and powering it up to see everything in action. Excited to see how it all comes together!
Regarding the NOTES….
Prototyping an accurate shooter unfortunately proves rather difficult, as the compression and aerodynamics of our own game piece will most likely differ radically from the official game pieces. Right now, we have to make due with these works of art:
Next Steps
Tomorrow, 3DM will be taking an in-depth look at potential robot concepts. We will provide you with some more footage of our kitbot, software and prototype progress. So stay tuned!
Written by:
@Bjorn - Lead Data Driven Decision Making
@Nigelientje - Lead Outreach
25 Likes
real
1 Like
We tested this as well, and the results were… bad. Hard frisbees get deflected and it works great, but soft rings squish with the impact and the energy of the shot is deadened. Not to mention that it starts tumbling in the air. The biggest angle deflection we could get before the shot was just entirely ruined was about 5 degrees, so you get very little for it.
I say this specifically because you guys might get different results with your custom rings that lead you down a path that doesn’t work with the real ones.
29 Likes
I really like your team’s idea for scoring amp/trap with the drainpipe. Do you have the CAD files for that design?
1 Like
Well that’s unfortunate… It makes sense though, given your argumentation.
This is exactly what we were afraid of when testing with our own rings.
Thanks for the feedback, I’ll pass it on to the appropriate sub-team!
6 Likes
We have uploaded the prototype on Grabcad: https://grabcad.com/library/team-rembrandts-2024-proto-drainpipe-1.
We didn’t test the concept since we still haven’t got notes. But when they arrive we will test them and share them!
6 Likes
Working Kitbot & Software
Once again, we’ve made a lot of progress yesterday with our software, and we also put our kitbot to action!
5 Notes Auto Path
We tried some initial auto paths with our 2023 robot to see what’s possible in a 15-second timeframe. They’re currently far from optimized, but it’s a good indication that we will be able to do more than 4 notes…
Shooter Deflection Flap
We tried out the deflection flap (*cough cough* a wooden plank) on our kitbot shooter (angled on wooden beams) to see if this concept has some potential. We mentioned this idea yesterday, but @Footie pointed out that in their tests it doesn’t quite work with the real game pieces since they are more sticky/compliant than our fake notes. Furthermore, we still want to test this concept later this week just to be sure, but it won’t have a high priority, since we first have to test other, more important parameters such as: RPM, spin, angle, wheel size, wheel material etc.
Working Kitbot
Today we have finished our SDS swerve version of the kitbot and tested it on our field. It works incredibly well, as seen in the video below. Of course, this isn’t something we will put on our final robot, because we’re currently developing better alternatives for shooting in the speaker. However, it’s a good learning point as we get extra insights on what’s already possible with a really simple robot, which helps us immensely with the decisions we have to make to stay true to our desired 70% robot.
Written by:
@Bjorn - Lead Data Driven Decision Making
28 Likes
maybe you have tried this, but we found that a pool doodle makes an ok NOTE substitute when formed into a 14 diameter ring. (the undersized ones are more similar)
Thanks for all your updates!
2 Likes
Can you share the CAD files and details for the other Field elements as well?
Yes good point. I tried some of the pool noodles we had already lying around in the shop, but their diameter was way above 2 inches.
1 Like
Y’all are flying. Really love seeing the auto paths being tested this early. That’s huge.
15 Likes
We printed out a TPU note and honestly it’s a pretty good approximation. It doesn’t squish as much when compressed as it would be in the kitbot shooter, but I think this was more of a feature of our wall count them anything else. Otherwise, it was pretty accurate, it felt very similar, and definitely has a use case if you don’t have real notes. I’ll send a picture later, but you can try this if you want something a bit closer to reality.
4 Likes
Pool noodles just aren’t dense enough to fly like the actual notes, and they aren’t quite as flexible. It kind of works, but I wouldn’t draw any conclusions based on them.
1 Like
For sure, they have many dissimilar properties. Surface finish is quite different too. Close enough for HP practice maybe?
Shoving a length of six gauge battery cable in the center hole of the pool noodle would help approximate the NOTE weight.
1 Like
