I’m curious how other teams has gotten better over the years that they have been around. Also what has worked and what hasn’t.
8 Likes
I don’t think I’m fully qualified to do detailed answers to this question…
But at a bare minimum, organization, structure, and mentor retention are all huge in my opinion.
Communication between all team members. Structure to know who does what, so there aren’t miscommunications.
That’s repetitive but I think it’s huge.
10 Likes
To add to what Eric has mentioned:
Protocols. For everything.
-Practice everything - have a list for field/pit/auto routines. Practice those protocols too in odd situations so you can iron out issues before comp. No silly mistakes costing matches. For context, we had no repairs outside of routine maintenance for 3 straight competitions including worlds. Largely because we broke it many times in our shop and found the problem areas early. Knowing when to stop making changes and let the dt test it is key. An untested bot on the field is a massive liability. This was the biggest change we made this year and it paid off in a huge way.
I’ll throw in that having an honest conversation about what your team can achieve and managing those resources effectively is a must. If you reach that goal, add more. Knowing your limits while pushing hard to exceed them is much better than trying to do it all and ending up with incomplete or untested bot.
Sometimes it’s tough to convince students and even harder to convince mentors but, imo, you will have a lot more fun with a robot that does one aspect of the game great then one that does it all ok. Being creative in your design phase and leaving opportunities for “extras” if you have time to build/program is a layer of skill that we are still trying to hone in on.
TLDR; practice and protocols
13 Likes
-
Practicing more regularly – We used to practice maybe once the robot finished (1-2 weeks) and otherwise just compete or do demos. Hearing a local team putting time in before off-season events changed to trying to have more after-school practice times.
-
Practicing with more purpose – Along with that just introducing more timed trials and other drills during practice helps
-
Simplifying mechanisms – Just have a couple of positions, using the field to line up closer to score, etc. In 2020 and 2022, we only shot close up. In 2023 we mainly picked from double substations. These choices may limit you to reach the highest competition levels but can raise your competitiveness.
-
Borrowing mechanisms – From older robots, from other teams, from COTS suppliers. The more you can do this the quicker you can finish. Our elevator was same as our 2018 elevator, our slider was the PWFusion slider (diy’ed), etc.
-
Buying swerve modules – we tried a custom holonomic drive in 2019 that was a learning point, and this year buying modules raised how competitive we could be by 5-10 spots. It kind of follows from (4).
-
Recognize the value of autonomous/end game – If you can finish the robot sooner, you can make strides here. There is often a good opportunity to gain advantages here, like being able to balance this year consistently would raise your team’s stock, especially at earlier events.
-
Being RP aware – Sort of goes also with (6) but thinking each match how the alliance will approach or meet bonus RPs can help to not miss ones, and help your rank out. In 2018 we didn’t work hard enough on 7 (and 6) and the one extra RP would have raised up the few spots to get us into a playoff alliance. So it is something we’ve been more aware of during quals since then.
15 Likes
After the COVID-19 pandemic, my team experienced a significant leap in competitiveness. Although I joined the team when it was already highly competitive, I’ve observed several factors that have contributed to our success. One crucial aspect is the effective transmission of information within the team. This ensures that mistakes are not repeated, creating a streamlined system. Additionally, we establish goals at the beginning of each season, not solely focused on robot success but also encompassing community impact and non-robot awards. Lastly, I believe that team culture plays a vital role. Setting high expectations for the team and fostering an environment where high school students genuinely enjoy spending time are often overlooked factors in achieving success.
2 Likes
We always mention continuous improvement. It is our motto, and though difficult, it is important.
While most teams want to do better each season compared to the last, knowing what you are lacking in and making efforts to improve it are critical. Perhaps for some top-tier teams it possible to be satisfied with one’s current state, but for us, there is always more we could be doing. There are better ways to build, technologies we aren’t aware of, more advanced programming techniques, and simply better practices.
Learning more is key, and without that additional knowledge it will be difficult to know what you can do to become more competitive or achieve your other goals.
7 Likes
I’m a huge proponent of using industry lean continous improvement tools in FRC. Not only does using these tools provide a path towards excellence, it also opens students to the world of process engineering, which might lead to a career for some.
Your robot and your students are both the product of your process. If you improve your process and reduce waste, you can increase the percentage of your activities during the season that are value added. In this case, lets call value “anything that leads to increased robot performance”. Another option would be to define value as “what increases students knowledge/interest in STEM careers” it really depends where your priorities lie.
Once you’ve defined value, you can conduct a value stream map where you analyze what parts of your current state build process are “value added” and which are not. Some of the non value added steps will have to stay because you can’t avoid them. Others are considered waste and should be eliminated.
After identifying sources of waste in your process, you can develop a future state process that avoids waste. I.e. “we discovered that we spent about 30 minutes every meeting looking for tools that had been misplaced, so next year through the use of shadow boarding and 5S techniques, we will reduce wasted time looking for tools.”
A comprehensive approach that outlines steps to increase value in your build process will lead to a complete product at an earlier date, which will allow you to test, iterate, and improve to competitive excellence.
There are plenty of other powerful lean tools that can be used in FRC. These included Root Cause Analysis techniques, Failure Modes Effects Analysis, PDCA (Demming cycle), Hoshin Kanri strategic deployment, barrier tracking, and even a simple point Kaizen philosophy. If you’re interested in process improvement tools, research the lean toolbox and see what you can find.
8 Likes
I’d look into 1678’s fall workshops, they post them on their channel and they do them every year. The one that helped the most for me is their strategic design talk.
Also, Getting your robot done early and having plenty of time to do drive practice helps to find out issues and improve how well your driver is able to play.
Unfortunately, I’d also say you kind of need swerve. Barring the possibility that FIRST implements anti-swerve terrain, you will need swerve to stay competitive in the future.
2 Likes
From a software side, the largest jump we made was getting PathPlanner up and running.
If you have not spent time getting some form of path following up, you should - it was by far the most impactful off-season project I’ve worked on and can be done with nothing but an old drivetrain and some carpet.
It’s not a silver bullet that will make your auto always work, but it was a huge part of getting our 3-piece autos up and running, which played a huge part in our wins this year.
4 Likes
Recognize the value of mentors.
30 Likes
pocket apples
13 Likes
How many losses were converted to wins because of this?
6 Likes
Adjust your team vision and goals. Change the culture. Its as simple as that.
5 Likes
If the goal here is to become “more competitive” I’d suggest a fanatical approach to game analysis and initial robot design. We decided on day one that we’d have a fool proof auto balancer and it paid off all season. We also in week one had a very spirited discussion on two competing design concepts with active debate pro and con. Once we’d settled in on it we had designated sub groups for various components…helped quite a bit by similarity to mechanisms from a couple of earlier seasons. And learn from your mistakes. Being top heavy the year before we built a machine with a ridiculously low COG.
Now there are other goals too. Don’t let things like ludicrous build hours interfere with learning or having fun.
5 Likes
2022 we ranked as the bottom of both of our district events and this year we were Alliance Captain at our first event and First pick of 3rd alliance at our second.
The things that we changed to get there were building within our means, if we don’t understand it easily, effectively and quickly we move to something we can implement effectively and iterate on that instead.
Next was driver practice, more practice = more better.
We also partnered with our schools PLTW program to train our students better in CAD and utilizing the teachers/mentors skill sets. We also started sharing resources such as CNC’s and 3d printers.
6 Likes
Programming:
The 2021 at-home challenges let us / forced us to figure out path planning, which as others have mentioned has been a huge boon for auto performance. With this added skill, low CG and planning for auto in the design phase is suddenly a main concern.
Parts organization:
Sometime around 2017 our storage space was getting to be a hot mess in a small space. Stuff in totes and cardboard boxes (or not even really in a container) and generally hard to find things. By 2020 we started using Stanley 10 and 25 compartment organizers, starting with electrical parts, so that things weren’t in bags several deep in KOP totes. This has let us know what we have, be able to find parts easier, and not take useless stuff to events, or find that we missed taking useful stuff (like bolts several years in a row).
In 2022 we redid our toolbox situation. Previously we had a single cart toolbox that we took to events and another that stayed at home. The tools on the cart were several deep and weren’t all needed field side, and the toolbox was annoyingly high on our cart. We got a smaller toolbox for the cart and a new one to stay in the pit which has let us spread out our tools to be (mostly) one layer deep, have “repair” tools field-side and “manufacturing” tools in the pit, and now our driver station can sit on our cart at a useful height and not be precarious.
Strategy:
In 2018 our analysis said to focus on the center teeter-totter, but we ended up making a really good set of 2007-style ramps for endgame and a mediocre elevator, basically ignoring the crux of our game analysis. We did okay-ish and qualified to states and worlds (barely), but mostly because other teams had little to no center Scale capability.
And in 2019, we spent like two-and-a-half weeks prototyping ideas without coming to great game conclusions, only to fall in the low or mid trap of scoring only hatch panels, climbing to L2, being mecanum, etc. We wanted to do something within our means (which we did, especially now that it was two weeks in), but the performance was disappointing. We did even worse than 2018, not even making it to States.
Going into 2020, the lesson of the past two years was: listen to your analysis, and decide in the first week to “do everything” and just go for it. Now that “do everything” doesn’t really mean “everything” for us, for example in 2020 we realized the Control Panel would be negligible in early events so we just ignored it (good call as at the Kingsford event there were like 2 whole matches that it was even eligible to be spun).
2020-2023 has earned us our first 3 winner banners and another finalist trophy, all as high seed alliance captains, and our first Innovation in Control and Autonomous awards.
As a strategy note for these games, our success has been substantially aided by being one of very few teams to be super consistent at the endgame ranking bonus (2020 level generator switch climb, 2022 hangar bonus, 2023 charge station auto+end game bonus). Consistency at these has been key at our first event of the season since the 4th ranking point has been rare and other teams just aren’t as consistent. So, in 2024, go dominate one of those ranking points.
3 Likes
Can’t stress it enough, build simple and finish early. For the last two years, our competition robot was completed and moving in 3-4 weeks. The rest of the time before our events is spent on drive practice, auto tuning, and iteration. The biggest overall change was really knowing our resources and building within our means. We’ve designed for the past few seasons with the idea that we shouldn’t, and don’t need to do everything. We limit ourselves to being really good at a few tasks and add capability later in the season as we see fit. Also, prioritize the RPs from day 1.
7 Likes
Another thing is build within means, simplify and get drivers practice before first event, and get programmers finishing a consistent autonomous before then. But don’t stop when the robot is handed over, keep designing and building either new iterations or features that will increase how capable and reliable your robot will be at your second event.
If you can incorporate this improvement potential in the original design even better, knowing where you can take the design. One phrase that sticks was something Greg told us about improving process, which was ‘Don’t major in the minors’ just meaning you should be focusing more on changes that improve your potential.
But all that work has to be after the your competition robot is ready for the first event. And for those only going to one event, you can get better by raising money for a second event (may be easier said). The ability to go to a second event with maybe some changes is invaluable.
5 Likes
High quality bumpers/spending time learning how to make good bumpers/learning how to properly budget build season time to make bumpers/figuring out how much material is needed for bumpers/etc. are all underappreciated priorities in my opinion. It really helps people learn the best processes for design and manufacturing.
6 Likes
Attention to detail goes a long way. Seemingly small details can have large effects and many small improvements add up to large benefits, often what separates the top teams from the rest.
Focusing on improving mechanism stiffness can go a long way (well, when the mechanisms is not intentionally compliant). Poor stiffness is one of the most common ways in which a mechanisms is unreliable. We’ve all seen those arms or shooters that aren’t very stiff and either break or are difficult and slow to line up. No amount of programming or driver skill will fix a mechanisms that physically won’t do what you want it to. Stiffer mechanisms have the capability of reliably doing what they are told to do, which is required for good programming and driving to have major benefits.
Many potential mechanical failures and small optimizations can be found while CADing when it is easy and fast to make changes instead of after a robot is built in which it is harder and slower to make changes. Many small improvements only take a few minutes at most to implement in CAD. Every potential failure that is eliminated in CAD is one less failure that needs to be debugged or fixed during comp or late in the build season.
6 Likes