Introducing the EWCP FRC Research Agenda

EWCP has developed an FRC Research Agenda: our attempt to coordinate a rough consensus on the next steps of tech development in our sport.

We invite collaborators to join us in refining the Agenda, which is in a preliminary state, and tackling the interesting unsolved problems listed in it.

EWCP FRC Research Agenda
Repository

Why we need a research agenda

A research agenda is needed to serve as a framework for intentional, organized study of the things we think are important-- the things that will make a beneficial impact in our competition community.

The EWCP FRC Research Agenda will reinvigorate our community’s interest in citizen-science at a unique point in the history of FRC: a time when teams are returning to their workshops and re-opening lines of inquiry that had been dormant during COVID. We also hope to inspire new researchers to explore these and related topics as they begin their exciting journeys in STEM.

Project goals

Build a rough consensus:

  • What is the current “cutting edge” of FRC?
  • What should be the focus of research if we wish to advance the cutting edge?

Support teams already doing good research and teams embarking on new high-priority research

Generally promote scientific thought, especially with regards to:

  • indigenous knowledge that may no longer represent the best practice
  • information originating from FIRST and FRC suppliers

Why we are qualified to present a research agenda

EWCP represents a diverse group of current and former FIRST adult participants with copious experience using our resources to grow the FRC knowledge base.

How you can contribute to the research agenda

We invite members of this community to help us improve the EWCP FRC Research Agenda. What items should be added or dropped? How should the items be prioritized?

31 Likes

I think having a competitive robotics experiments youtube channel could be an interesting concept: for example, in FTC land, while people do theoretical calculations/estimates of how different drivetrain configurations may perform (the general consensus is that a drive ratio of 13.7:1 to 20:1 on the legal 550 motors would be optimal), there is little empirical data on how they compare to each other. Something of this concept would be something interesting to explore, although it would require a decent amount of up-front capital (likely from existing teams).

4 Likes

What type of teams are you hoping to help with this effort? I’m struggling to parse the goals exactly, but by trying to increase the cutting edge of FRC you’re leaving most teams in the dust.

Browsing the topics here they’re mostly in the optimization realm, and aren’t going to help a team make the jump from missing elims to playing after lunch.

The chronic issue with most teams is attempting to build past their means and/or not working efficiently enough to show up ready.

If the goal is to help teams move from finalists to champions, I guess this makes sense… But if it’s for the majority of teams I’d recommend focusing on more approachable topics.

31 Likes

I’m not the leader on this project, but personally speaking I think it’s a good way to answer some questions that have always plagued some group chats with “Is X true? Wait, I thought it was only true in this case?” etc. There’s been a lot of conflicting information put out by different teams over the years on various topics.

For example, does radio placement actually matter? Do breakers actually degrade over a long period of time?

There are some less serious ones like predicting GDC responses…

I’ll be chipping away at ideas relating to automating pre-match checklists. I know 254 has done it in the past and many many other teams; if anyone has any feedback on this topic I’m more than happy to listen.

10 Likes

A pretty big topic for me while I was a student in FRC was passing down software knowledge effectively. I’ve heard countless stories of teams starting almost entirely from scratch with their code because they had one programmer who graduated without training a replacement. I think taking a look at good practices for training new team members in software and passing down knowledge could be helpful, especially if it focused on how teams without heavy software mentorship could use peer learning to pass down knowledge from generation to generation when their students graduate.

I think my approach where I spent my senior year teaching the rookies and helping our upperclassmen write the code for our robot instead of doing it myself mostly worked out, but Covid did throw a wrench in my plans and we can only wait and see how future members of the team do.

7 Likes

Increasing the cutting edge doesn’t help many teams short term, but those innovations can eventually benefit a much larger number of teams down the road. Take trajectory planning for example, which was pretty cutting edge not that long ago and was only used by a few teams. Those few teams and some developers found a way to package it in a more approachable form, and now we have tools like pathweaver which are very accessible and commonly used.

8 Likes

I’ve always liked the idea of essentially having an FRC equivalent to Linus Tech Tips, for lack of a better comparison. Rigorous testing of new parts and components done in time for teams to get their hands on them, and general experiment videos between releases. If it’s not video content, it’d at least be nice if someone independent from any supplier or team could get ahold of them, get some testing done to get specs and test them out in some common applications to see how they perform.

  • THE “NOT FOR DRIVETRAIN USE” DRIVETRAIN
  • OLD HARDWARE, NEW ROBOT (thumbnail is the channel’s host with a shocked expression on their face while holding up a globe motor)
  • BUILDING FINAL 2011 DRIVETRAIN!
  • FMS Show #486: RFID: Is it a GOOD idea or a BAD idea?

As for the actual topic of the thread, I think what’s laid out is going to be pretty helpful. I’m especially interested in the meta-FRC stuff near the end, and I have a few ideas for that:

  • What measures most often result in a safety award and do they make competitions safer?
  • Is Chairman’s equitable at the regional/district level? Can a low-resource team win with local outreach alone or do you have to do something big?
  • Is it possible for there to be a universal robot performance statistic?
  • Are further events similar to the 2017 Festival of Champions viable? Is there a way forward for a culminating event after the first FOC was largely underwhelming for FIRST, and what changes would need to be made to make it worthwhile?

On the more technical side:

  • What exactly is the source of the magnetic encoder noise described in the FRC Characterization guide, and which magnetic encoders specifically are susceptible?
  • Which brand of battery is objectively the best for FRC? Community consensus is MK, but has this been tested empirically?

I’d also like to know, who exactly is involved with EWCP? I assume this isn’t intended to be a mystery anonymous group since I’m pretty sure everyone involved has said so publicly, but it would be nice to have it all in one place since I honestly just thought this was a Fantasy FIRST team until now.

7 Likes

I’d agree with that, but I’m not sure what “research” could be done there – there are no real unanswered questions with regards to what prevents teams from being good. Hell, you said it yourself:

That being said, we’re working on something regarding the latter, since our work process has been abysmally inefficient for a while, and we’ve been working on things like space layout, organization, and design for assembly, though it’s going to be a while until we’re at a spot where I’d feel we’re a good example for other teams to reference (speaking of which, if anyone who sees this has any advice to build more efficiently please shoot me a private message, we’re always looking for new ideas)

2 Likes

The Blue Alliance has you covered!

Sorry @Greg_Marra @Tom_Bottiglieri

26 Likes

EWCP - About would be a solid place to start.

I coulda sworn we had a History page at one point, guess that’s a todo.

Short version - a group of folks got together a while ago to talk robots and build more competitive ones. Then that went off and did a podcast for a few years, a few blogs, and most recently a scholarship. This is the latest effort to help push towards a world with better STEM opportunities.

As a general rule, with the exception of board of directors, we haven’t published member lists. Not out of any intention to be secretive just simply because I’m not sure there’s a point.

2 Likes

just looking through the website, I noticed that the wiki requires a login to access it – is it still under development?

1 Like

I’ve always wanted to build an instrumented dyno…

4 Likes

Good criticism! The project is meant to support efforts to uncover knowledge that, once uncovered, will better equip all teams. The project is not directly meant to help teams be competitive in the near term.

Some good examples from FRC history of things that have advanced the field, despite not immediately helping most teams play after lunch, include:

  • Chris Hibner’s “How to make sure your robot will turn” Ian MacKenzie’s Drive Train Kinematics
  • Ether’s kinematic derivations of swerve and omnidirectional control theory
  • 234’s drivetrain testing
  • 2363’s measurement of the adder for exact c-c #25 chain spacing

It’s a good point that we should prioritize agenda items that have the potential to help most teams. This is the major reason to develop a research agenda: to create a forum where we can rank certain lines of inquiry above others. We hope to flesh out each item with a description of

  • what the actual problem statement is
  • why a solution is desired
  • who it will benefit and how
  • why it’s possible to solve now

An apparent focus on optimization might show our limited creativity in developing the research agenda so far. Or… (And I dont think this is the case) it might show that there’s generally less unsolved problems out there. The examples I gave above (from a landmark 2003 whitepaper to a 2018 CD post) shows a trend of narrowing scope and a shift of focus from theory to empirical measurement.

6 Likes

I think this is THE fundamental example of a paper that did immediately help teams play after lunch. It wasn’t bleeding edge teams that weren’t turning.

1 Like

I’ll disagree on both counts…

  1. “How to make sure your robot will turn” was just the memorable alternate title of a paper actually called Drive Train Basics, which established a baseline understanding of the traction forces on a drivetrain.
  2. Einstein teams were struggling to turn in those days. E.g. 469 2003 in this semifinals match: https://youtu.be/vnwl31zoAPI

My point to Adam is that research that advances our community does not necessarily need to be immediately helpful to most teams. If it does, that’s great!

2 Likes
  1. Regardless of which title you prefer, the paper was aimed at helping increase a fundamental knowledge of drivetrains. It was mostly effective at helping borderline teams, such as my own, create more effective drivetrain solutions.
  2. 469 didn’t “struggle to turn” because of drivetrain issues. Watch their turn in autonomous, it was butter smooth. Their driving in tele-op is the result game design and strategy. They’re on the red side of the field, trying to defend the red stacks without pushing any totes out of the red zone. 341 flips, and 175 is initially on the blue side. 469 begins crossing over the ramp to blue when 175 comes under the bar to red. 469 attempts to return to red to save their stacks, but navigating thru the field of crates is not easy work. 175 doesn’t have to particularly care about navigating carefully thru red, as their goal is to descore totes. The game design that year inherently led to teams have to maneuver apprehensively on their side of the field. Visibility was also a concern (particularly on the far side of fields). To steal a phrase from 341, “we knew where we were by watching the trail of destruction among totes on the far side.”

If you want to reference a paper from roughly that era that was aimed more at bleeding edge teams, Ian MacKenzie’s Drive Train Kinematics would be a far better example.

1 Like

I don’t know what the right answer is to help lower performing teams (but I’ve been trying to figure it out)… Much easier to say what it isn’t.

Cool stuff for the sake of science, engineering and/or its own sake is cool, but isn’t it.

  • Chris Hibner’s “How to make sure your robot will turn” Ian MacKenzie’s Drive Train Kinematics
  • Ether’s kinematic derivations of swerve and omnidirectional control theory
  • 234’s drivetrain testing
  • 2363’s measurement of the adder for exact c-c #25 chain spacing

Agreed on the impact of the first and how it lead to the popularization of the various standards of 6wd’s, but I don’t think the rest of had much of an impact on most teams.

Blockquote Increasing the cutting edge doesn’t help many teams short term, but those innovations can eventually benefit a much larger number of teams down the road. Take trajectory planning for example, which was pretty cutting edge not that long ago and was only used by a few teams. Those few teams and some developers found a way to package it in a more approachable form, and now we have tools like pathweaver which are very accessible and commonly used.

I have an unscientific estimate that less than 15% (edited up from 5% after some very loose math) of teams effectively use the paths.

3 Likes

My unscientific estimate is that this season meant we finally had enough programming time to make use of them (after past seasons of being disappointed by other trajectory things not working out for us).

2 Likes

My even more unscientific estimate is that usage has increased a good bit higher than 5% given that almost weekly/daily there is some active thread where a team is trying to use path planning/frc-char. I assume by the end of the season they will be “effectively use paths”.

1 Like

15% translates to hundreds of teams, which is far more than the handful of teams that were using it a few seasons ago. I’d expect that number to continue to increase in future seasons too.

1 Like