Aerial Assist--Perspective from a low-resource team

FIRST General Forum
1

Well, this year, I’m with a new team (well 2, but it’s complicated), and I’ve spent only the last week working with them. It’s a great team with awesome people, but in all their history (5 years), they’ve only gone to eliminations once, and it was as a pure “box on wheels” role to play defense. The team has good ideas and can implement them, it’s just that something never seems to go just right for them to be successful. This year however, they just got more support to make a limited number (think like two gearbox plates total) of CNC parts. Their shop has 1 drill press, 1 pretty crappy mini lathe, and a small bandsaw + drills/hand tools. I’m used to having 3 mills, an awesome huge CNC lathe, CNC mills, a CNC router, and all the tooling in the world to prototype with, so I expected this season to seem, well, a little slow.

I was incredibly surprised when on Tuesday, the first time I met with the team, they’d put together the kitbot (they’ve got some pretty great students, some of the best I’ve ever worked with), and a few kids had put together an awesome pickup device, that, due to nifty angled rollers, could pickup the ball from any angle in the front, and worked better that the Ri3D teams did.

A few days later, me and one of the kids put together a catapult with a clever joint I’ve nicknamed the knee. In the unloaded position, it’s like a bent human knee. you push on it with a cylinder, and it pushes up on the end of the catapult to load it. As it gets closer to the end, it gains mechanical advantage due to the nifty angles, then goes past perfectly straight and locks into place, no latch required. Then, a little string pulls back on the knee, lets it go and the ball flies into the goal. The whole thing took a student and me less than 5 hours to build.

Right now, we’ve got a robot that can pickup a ball pretty well, shoot it very accurately (we use the edge of the low goal to line up), and catch it’s own truss pass, using the kit bot, pneumatics from last year KoP, and first choice items. It’s the end of week one, and the robot’s done! This team doesn’t usually start auto mode programming until after stop build, so we don’t know what to work on.

In some ways, it’s a blessing for low resource teams, but I’m imagining my old team right now, knowing that they’ve probably already assembled their custom chassis, built their shooter/pickup already, and are wondering what to do just like us. I remember being blown away by the robonauts being able to have a functional robot by the end of week three, but now, a low budget, really simple team able to meet our goal of “do everything” at the end of week one?

So, as my team has asked me to ask CD, what next? What innovative features has your team come up with to make their robots better, and how can a team, like mine right now, build something else? We’ve already got a launcher/pickup/blocker and we don’t know what to do next. We’ve considered a new drive (swerve or just a fast, shifting 6wd), but we can’t afford it or justify it. The kit drive is pretty darn good.

2

This team doesn’t usually start auto mode programming until after stop build, so we don’t know what to work on.

You could get started early on that :slight_smile: Missing in autonomous will be one of the worst things that can happen, score-wise (the worst I think being hitting the truss on a Truss Score and having it bounce back past your robot)

3

Do some testing,

  1. Run your robot’s shooter many times, see if anything shows wear and then upgrade.
  2. Test repeatability of robot movement, start it in the same place, and then move forward, say 10 feet, run it several times, does it stop in the same place every time?
  3. Come up with other things to test for accuracy, repeatibility and reliability.
4

First off: Congratulations! That is an incredible feat for any team.

Work on code, get some sensors in there and automate some tasks (does the intake need to be in a particular position to shoot, and a different one to collect? Code it to one button, rather than making the operator remember). Get some auto code, develop multiple auto routines. Can you make it do 2 balls in auto?

Practice, practice, practice. See if there’s a team nearby that has a full field, go test it out. Find out what the limitations of your machine is. Get a lot of run-time on the robot and find out what wears down or breaks. I’m sure you’ll find stuff that you can improve.

Develop strategies for various situations. Can you shoot over a 60" tall defender? If not, figure out how you maneuver to avoid them. What’s the easiest portion of the field or robot orientation to pick up a ball? The hardest? Train your human player on how to quickly and efficiently load the robot.

You can get to your first event with not only a working machine, but also with a team and drivers who know the robots strengths and weakness and how to play to those. You will be a more effective alliance partner because you can coordinate with your alliances on how to best leverage your robot’s capabilities.

Also, make a killer reveal video. You’ve got 5 weeks!

5

I have never seen a team with a consistant and successful double autonomous mode not Make it into Elims at a regional so you may want to start working on that.

Driver practice - driver practice - driver practice. Find other teams that have an 08 robot that can play this game or are close to where you are at and practice with them.

6

If you have the time and resources the build a second robot with the knowledge learned from the first. The first one then becomes a test/practice bot for the drivers and programmers to work on.

7

The best thing I would say to do is practice driving. Looking at the game, many of the matches will be decided by who has the better driver.

Also, building a second bot would also be benificial so you can continue to practice driving after the build season is over.

Autonomous will also be huge considering that since many of the teams (from what I have seen) will be decent or better at scoring during tele-op, so matches might also come down to which alliance missed an auto shot. So getting an early start at auto is always a good thing.

8
  • Build a better shooter so you can shoot from a less defensible position.
  • Build a pickup mechanism which can withstand being hit by a defensive robot driving at full speed.
  • Build bumpers which can be color swapped in 15 seconds.
  • Build a robot cart with tool storage.
  • Build a practice bot, and do drive team practice.
  • Redo your robot wiring so any component can be replaced in 15 seconds, and anyone who sees it thinks you should win a quality award.
  • Build an amazing pit.
  • Find another team and help them.
9

Thanks for the suggestions, I’m trying to convince the team that auto mode programming should start. Trust me, we’re not really done. We have to do driver training, figure out the accuracy and improve our shooter, we need to have a rock solid auto mode… the list is endless.

However, the team has a huge mechanical team that’s disappointed at not being able to make more parts. They want to learn new stuff. That’s why last year, they decided to build a 30 point climber (which took 55 seconds, 5 seconds under goal), even though they knew it wouldn’t be super competitive, just because it had a lot of cool mechanical aspects. That’s what I really love about this team., They get it. The students (and mentors) are really inspired to spend hour after hour working on every little detail in order to build cool stuff. They aren’t driven by winning. For this team, the competition is only a really small part of the robot season. It’s all about working together in the shop making neat stuff, learning, and having a good time.

tl;dr – we know we are far from done, we’re just looking for stuff to build/design, and we really like most of the game.

10

Driver practice is important, you should do that whenever you can. You can get your driver station set up, and have the drivers work with the programmers to set up controls to the drivers’ liking. Also, programming an auto mode is much easier when you have unrestricted access to the robot.

11

That’s great! I agree with everyone else, look at the game, strategy will be vital so spend time working on yours, driver practice, and a lot of it! Improve and debug everything you can, break everything early so you can know how to fix it and where week points are. See what you can make better! Great job to your team!!

12

I like the idea of the Knee joint. The only thing I would be worried about is interference from another robot and the joint being jostled into firing unintentionally.

You didn’t mention it in your post, but CAD gives your team a good way to keep everything documented instead of just hand drawings. It may be too late for this season to use it, but it’s very useful for a team. It lets one check out the geometry without having to be in the robotics room.

I love the fun attitude you guys have to the game, but keep in mind, part of the challenge is to figure out the best way to play the game, theory crafting. Like how many cycles can we get off, is it worth doing this if it only gives us so many points. One of the mentors who just came into our club is there to replace our chief strategic mentor. [The new mentor is a manager at a contracting corp, he gave the past mentor a job. That job led to another one in Oregon, across the country. The new mentor felt he should fill the role since he helped move the other out of the local ares].

To each team their own, but I know many on my team who love analyzing the game to no end and finding an optimal way to win. If you guys just want to build crazy stuff, that’s all good too. But why not both? haha

Do you have a link to your 30pt climber, i’d love to see it?

13

So, as my team has asked me to ask CD, what next? What innovative features has your team come up with to make their robots better, and how can a team, like mine right now, build something else? We’ve already got a launcher/pickup/blocker and we don’t know what to do next. We’ve considered a new drive (swerve or just a fast, shifting 6wd), but we can’t afford it or justify it. The kit drive is pretty darn good.

You design your shooter to have variable range/arcs so that you could shoot more vertical for the truss and a flat trajectory for scoring in the goal. Just be careful that your “improvements” don’t make the robot lose its current functionality.

14

Congrats!!! It’s really awesome that yall are making sure great progress!!
Here’s my input on what yall you be doing
Yall might want to have driver tryouts since yall have the extra time. By doing tryouts, you know who you want on your drive team, and then you have them practice until the wheels fall off of the robot:D
Don’t stop brainstorming! Every time your team is together, have them take another crack at bringing up a new idea to make the bot even better.
I hope yall do well!! It’s always promising to see other teams so far along in the process

15

Dear Mr. magnets,

Upon reading your story I could not help but draw some similarities to the team I came from (and now mentor). Now I want to share my perspective as well.

My high school sophomore year ('03-'04) was the season the team finally moved out of our founding sponsor’s basement and into a new build space. A decade later our shop has a mass of hand tools, 2 drill presses, a vertical and horizontal band saw, a chop saw, a circular saw (for our practice field), a vice, an air powered rivet gun and a compressor to operate it.

Though what we can create in shop sounds very limited, most of the time it’s all we ever need! If it’s not enough we have the support of our engineering mentors who have access to better equipment at work. As an FRC alumnus my experience has shown me that it’s not what tools you have, it’s how well you use them. Our last 2 drive bases were nothing but 1"x1" square aluminum tubing held together by pop rivets. Heck our 6-wheel tank drive from last season was one of/the best drive train I have seen from our students.

It also sounds like the students you mentor understand how important the kit bot can be. With a drive base assembled and running in less than a day, they can put their focus on the other features of their machine. The coaches and adult mentors of the Enforcers agree, if students learn and apply the engineering process, it won’t matter how well your robot performs in the end. They’ll have gained experience to use for future seasons and beyond. Seeing this process in action today is a big reason why I stick around to mentor the team that gave me so much valued wisdom.

In a nutshell, if the students put forth the effort, it will always be worth the failures. Make sure they keep at it. :]

16

First of all- color me impressed that you’re done at such an early rate. Using the kitbot and being smart about your mechanisms like your team seems to be doing is a great way to finish early and get in driver practice.

I would recommend you try to find some practice space and really test what you can do. How fast can you complete a cycle on your own?
Build two more kitbots to practice doing assists with if you have the resources.
See the exact range of your mechanisms, see what they can do that might be unexpected.

Last year we built our robot to shoot 3’s from in front of the pyramid, we discovered through some practicing and trying different things that when our shooter was in the “up” retracted position, we could shoot into the pyramid goal, which gave us another option when cycling under heavy defense.

Really knowing your robot’s capabilities inside and out can give you a HUGE lead on the competition at your first regional, and even throughout the season.

Maybe improve your mechanisms further. Fine-tune the shooter to get a larger sweetspot, improve your catching abilities, etc.

I hope you guys have an awesome season, because you seem to be off to a great start.

17

This should keep your mechanical team happy:

  1. Drive it like you stole it until it breaks*
  2. Do a failure analysis, redesign/remanufacture the broken component, or put into place a procedure that will detect the failure pre-match.
  3. Goto 1.

You’re probably correct that many other teams are at the same spot you are, and will now be iterating their designs to make them faster, more reliable, and more effective. Simply having a robot that can do the game task won’t be enough. You’ve got a fantastic chance to build a 6th or 7th version of your bot by ship day. That should keep your mechanical guys busy.

*“Breaks” and the associated fixes can mean many things:

  1. Did the driver screw up? Maybe your control software is too confusing and needs adjustment.
  2. Did the robot munch its own part? Maybe you need sensors to prevent that from happening, or you should design the part to be less likely to wreck itself
  3. Did a part simply break? Time for a redesign or beefing-up of that part.
  4. And of course, you can simply redefine success. Can your robot pick the ball up 10 times in a minute? Change your requirement to 15, and see what needs to be done (driver training? pickup redesign? different ball-approach tactics?) to get your effectiveness up to that level.
  5. Did a normal wear component break? How easy was it to replace? Could you design the robot to have greater access to that part?
  6. Was something forgotten before a test session and delayed the start of the driving? Make some checklists or maybe even have the robot software self-check that everything is present and working.
  7. Are there areas of the field that your robot won’t be able to pick up the ball, like corners or in the open field? You could change the pickup to make it more effective in those areas. Also, see [4].
  8. Can your catcher catch a ball thrown blindly (a human could throw it backwards over their head) 100% of the time?
18

Our team works our a classroom and have very similar situations regarding machining. What we have learned is that we need to build within our limits but still somewhat custom just takes careful practice. We manage to have a decent robot, at least from the past couple years. We love to bend 1x1 tubing for framing and mechanisms…its very hard but takes very little resources :slight_smile: btw, a good 2 axis vise for a drill press makes a world of difference!