My team and I were throwing around the idea of building two identical robots. The first being a prototype and the second being the competition bot. The reason for this would be to actually have time to practice driving and finishing up code during the period between bag date and the first competition. I’ve heard of other teams doing this but I have never heard the results. What are your guys’s thoughts on this, any flaws, suggestions, or tips? Thanks.
We went from placing 52nd at our first competition (Iowa Regional) to placing 5th at our second (Seven Rivers), primarily due to additional practice and autonomous refinements enabled by our 2nd robot.
The downside: it’s a lot of extra effort. We started much too late in my opinion, as it took us nearly 3 weeks to complete, so we had barely any practice at our first competition.
I think you’d find that most of the top tier teams do some version of this. Some, like ours, build a prototype robot and then take what they learn from that and build the competition robot. We do everything we can to make the two software compatable so software can continue working after bag day.
Our prototype is usually about 75% as good as the competition robot and breaks more often but we learn a tremendous amount from it. Our drivers get invaluable drive practice on the prototype as well.
Other elite teams seem to somehow create two identical competition-ready robots. I assume there is a prototype stage somewhere before those, at least at the subsystem level. Maybe they can speak to that.
On my time on 20, we would build two (mostly) identical robots in (mostly) parallel. This had lots of benefits that we took advantage of:
- Autonomous refinements, as previously stated.
- Breakage, as previously stated. In particular, this came in handy in 2014 when our drivetrain bearings blew repeatedly. Our ability to replace failed bearings improved drastically.
- The ability to swap robot parts in build season. While the idea was that the first make of a part would end up on the first robot, there were times when we were getting close to the end of the build season and needed to use finished assemblies from our practice bot on our competition robot. We didn’t ever pull major assemblies from the practice bot to be our withholding allowance, but we did use our practice bot to create and test a FCS blocker in 2013 and improve our catcher panels in 2014.
As far as prototyping things, they were usually done out of wood with the occasional lexan component. The geometries from the subsystems were then put in CAD and handoffs between subsystems created. The goal was to have this all done about a week from the kickoff broadcast. Some teams create a mock of the robot in wood for drive practice while the real robot is being manufactured. That’s something I’d like to try someday.
I concur with the previous posters - a practice robot is a lot of extra effort, and not just in terms of build time. There’s lots of extra resources that get thrown at it – more motor controllers, a second RIO, etc., in addition to the parts you fabricate. Furthermore, to take full advantage, both robots should be built to CAD. If there’s too much variation, then the parts aren’t interchangeable.
FWIW, 6844 isn’t doing a practice robot because the Utah regional is week 1 and is our only event this season, unless we qualify four Houston. The best use of our time and building resources is to build a solid bot and have lots of driver practice. Furthermore, as a rookie team, funding and resources are being spent on building up tools for our shop and other things. That might get reallocated towards a practice bot in future years, along with a second regional.
I would recommend analyzing a practice robot in the context of Karthik’s strategy presentation. Because building a practice robot takes resources, the golden rules and the remarks on tradeoffs apply.
Pretty much what Dale said. Best way to keep going after bag. Besides getting more development and durability time, the drivers get a ton of practice. Helps to have a field.
Downside: consumes time and money, the 2 things teams have the least of.
3925 builds two robots, we don’t quite have the manpower and time to make them identical but it still works out for us great! All the benefits mentioned have been true for us however the most valuable in my opinion is the advantage of being able to make upgrades in between or before competition.
Last year we were able to redesign our whole superstructure after our first comp and add a shooter to our robot before worlds.
Team 291 has done it the past 2 years and have found there is a very wrong way to do two robots, and a right way to do it.
In 2016 we did it completely wrong. We are a relatively small team with average resources, and we overextended ourselves. We made two exact copies which was completely unnecessary. We wasted time building two climbers when we didn’t really need to practice with it after bag day. We tried to do too many things, and although we fine tuned the auto it didn’t make it worth it. Our driver also didn’t practice enough to make it worth it. He only got around 2 hours of drive practice in the whole of post bag :mad:. We did horribly at both competitions, all of our mechanisms were bad, and we didn’t make playoffs.
2017 was completely different. We took the lessons we learned the previous year and applied them. During our strategy sessions on kickoff we decided that the best way for us to go was to climb and do 5 gears per match. Climb and do gears and be hella good at it. We decided we would make a relatively simple, reliable, extremely fast (19 fps) robot, and give the drivers tons of time to practice. Not even touching fuel. We also decided the practice bot didn’t need a climber. The only thing the drivers needed to practice was lining up, and it wasn’t used in auto, so the practice robot didn’t need it.
We ended up having a fully working robot in week 3, and it was completely ready for competition in week 5. The practice robot was finished in week 5 as well. This gave the driver (me) a ton of time to practice and work on auto. We also were able to attend a week zero comp, which was very enlightening. We found tons of problems with the robot that only surface after lots of use, and we were able to fix them at competition before they even happened! I, as the driver, practiced at least 1.5 hours per day 4 days a week!
I was able to practice lining up with the peg over and over again, hundreds of times so I was able to do this at competition! That to us was totally worth all of the extra hours invested in a practice robot. We achieved our goal of getting 5 gears incredibly reliably, our record is 7 gears. Mind you we did not have a floor intake, we only loaded from the chute. We had a great auto, and a great season. We made semifinals in both competitions, ranking 4th at Buckeye. We were finalists at the WOW championship, and won both the Steel City Showdown and the Mahoning Valley Robotics Competition offseasons!!
So in short, practice robots can make all of the difference. I think we owe much of our success in 2017 to having one! But don’t over extend yourself! Be smart about assessing your own abilities. We had a relatively simple robot this year which allowed us to finish on time and iterate to optimize everything.
Another benefit of building two is that the practice robot is a great way for an upcoming technical lead to learn the ropes! The practice robot is less pressure for an upcoming leader, but it still is very challenging! 2nd robot lead is essentially the feeder to Chief Technical Lead on our team.
Try to avoid any differences between robots that require code changes to connectivity to sensors and actuators.
Definitely snapshot (version control) the code you had working on bag n tag day. That might come in handy later to help you figure out what changed after bag n tag day.
Do expect that you will need to recalibrate on your competition robot (at competition) because of code changes on the practice bot. Plan for it.
Enjoy having many extra weeks to improve autonomous software.
Building two robots is so old fashioned these days. The cool kids are building four…
Nah dude, the cool kids are absusing withholding allowance and unbag windows to build their robots and spending those six weeks strategizing and refining a design.
The bag is dead, long live the bag.
We at the Cyborg Cats have built two robots since our second season, and CNCs definitely help out. Over the years they have become more and more identical, and it takes meticulous documentation to keep them straight. If you can afford it, go for it. If you can’t, go raise more money so you can.
Pros:
- Driver’s practice and Autonomous practice (if only we could afford a third robot) because a good driver is better than a good robot
- We can make a whole new system on the other robot and use it for the withholding allowance
- More machining practice for us students
- Pit crew can practice assembly/disassembly
Cons:
- A lot of extra work after an already tremendously heavy workload
- Cost is great
Here’s my take on the matter as a student:
Last year, we built 2 robots when we only had about 5~6 people working on the robot even during build season. Of course, it was a lot of effort, but was it worth it? I would say a lot of the benefits come from the construction method that you employ. We did sheet metal construction last year, a la 148. It strongly hampered our development of new mechanisms seeing as sheet metal construction lends itself to strongly integrated subsystems. As such, we couldn’t find a reasonable way of adding functionality to our robot without extensive changes to the frame. Yet, it also had its benefits: driver practice. As many people here have already said, driver practice is one of the biggest things that determines performance in competition. We may not have improved too much in terms of ranking at our second regional last year, but I immediately noticed our driver performing much better and much more comfortably.
In the end, I believe that given a reasonably modular robot design, building a practice robot is definitely worth it, given that you effectively utilize it for the development of new mechanisms or driver practice. Even for driver practice, building a practice drivetrain (with some added weights) is sometimes all you need. In addition, your programmers will love you for being able to code autonomous in between regionals.
Just my 2 cents.
If you have the resources (money, time, and people)to both build a second robot **and **continue drive practice after bag day, there’s definitely benefit! 3946 started down the two robot track in our third year - our first robot was clearly a prototype, as it featured a frame made of 2x4’s and 2x2’s, with a few motors and such sticking outside the frame perimeter. The next year, we started building “twin” robots after the prototyping was done. This past year, we had two rolling prototype chassis as well as two nearly identical robots - we made improvements to our gear hanger after bagging which we transferred to the competition robot as part of our withholding allowance.
The key things to be gained from “extra” robots are:
- Drive practice, particularly after bag day.
- Continued improvements based on lessons learned mostly from drive practice.
Each team must individually decide whether these gains exceed any other opportunities lost as additional effort and money goes into building and possibly redesign.
I would be interested to hear about the production schedule that teams follow when manufacturing a second(or even third) bot. Is it preferable to build the set in parallel, or focus on completing the competition machine first? I suppose it would depend on a team’s resources, but I would still like to hear about other team’s methods.
There’s two schools of thought. Building them in serial lets you get one round of iteration in during build season. Design/build the practice bot, see what works and what doesn’t, make some slight design changes for comp bot. Update the practice bot to match if you have time left, or immediately after bag day.
The other school of thought is to build them in parallel to ensure that they are as identical as possible. This prioritizes practice robot as a software testing platform/tool for driver practice. This is the route I’ve personally done in the past.
We’ve recently seen some teams actually build 3 robots, where the first robot resembles more of a prototype robot, followed by identical practice and competition robots.
You will need to decide if you want to build a prototype and a competition robot, or two identical robots. There are some differences.
If you go the Prototype / Competition, you start building earlier. You get to what you think will be your design and build one robot to that design. Then practice, experiment, break, refine etc. When you are done refining and know the time is ending, then you take all the lessons and improvements into the competition design. This is the approach we take, and then we try to go back and make any significant changes to the prototype to make it closer to the competition robot.
If you want to build two identical robots, you will do more prototyping and refinement of the design before you commit. Then you can make two of everything and build two robots that are very close to the same. Many teams do this are are very successful at it.
There are good and challenges with either approach.
Another advantage of 2 robots is off-season event. You can usually enter 2 robots, and that gives more people drive / drive team experience during the summer and fall, helping to prepare for the next season.
Synopsis of 3946’s schedule, 2017:
- Week 1: Read, strategize, prototype, basic drivetrain parameters
- Week 2: Build drive train 1, continue prototyping/eliminating manipulators
- Week 3: Program/test drive train 1, build drive train 2, design manipulators, purchases
- Week 4: Build & test primary manipulator(s), develop check lists
- Week 5: Drive practice, secondary manipulator(s), develop scouting forms
- Week 6: Drive practice, complete second robot, update first robot, autonomous, iterate
- Week 7: Drive, decide which to bag, bag, and iterate
- Week 8+: drive & pit practice, scouting, iteration, competition
Prototyping includes standalone devices, and those on rolling/driven platforms as needed.
We have the same group of people build the second of each item shortly following the first. This allows some improvements to take place and shakes out issues we may have missed in CAD. The drive building group usually disperses to form the core of the secondary manipulator group and drive teams; pit is a cross section of the build subteams. The past few years, we have mostly competed with the second complete robot and practiced with the first, back-fitted robot.
Edit: results of 2017: We built both (KOP with mods) drive platforms (plus a third with no motors for off-platform builds) ahead of schedule. We built one prototype robot with a competition drive train but passive hanger to start drive practice on schedule. We finished the second with an active hanger and climber a bit late, and re-built the first. After bagging, we added a pop-up extension to our gear intake slide and improved the active hanger on the practice robot, which we transferred to the competition robot at competition. We also brought our practice robot’s climber for cheesecake (we used parts, not the whole thing). Both drive and build felt they had too little time; we’ll try to make some hard decisions (dropping things) a bit earlier this year, and some other minor tweaks.
This season, we built 2 robots–a competition bot and practice bot. After the competition bot was bagged up, we started adding things to the practice bot and prototyping new parts. For example, our initial hopper design wasn’t the greatest of hopper floors and barely fed balls into the shooter properly. Aiming to be somewhat of a fuel bot by the time the NYC regional showed up, we constantly redesigned the hopper. If it weren’t for having a practice bot that we could experiment with, we wouldn’t have a great hopper floor.
Drive team can have lots of practice before competition, and programmers can program new autos and improve shooter accuracy. Though it is a lot of extra work to create a second robot, it ultimately pays off.
This is always a fun question this time of year.
A good way to start navigating which way you go is to look back at previous build seasons and see where your robots have been when you put them in the bag. A change of plans to make multiple machines can quickly bog the build season down and slow the overall process down to where teams would have been better only with one. I’d argue we would have been slightly more prepared for GSD this year had we not tried to build a full second machine before we scaled it back to just a drivebase and in the end it never went further than a welded chassis.
Building two (or more) robot is doable, but like many have said your biggest struggling point will be time and money. Alternatives can include building only a second drivebase so your programmers can work on code, drivers can practice, and you can withold a mechanism easily to tune and practice after bag day. This can cut down on a lot of time and money spent duplicating the parts in each subsystem. 3467 went this route in 2013 since we never had more than one complete iteration that year but the practice chassis was a development base after bag day for a big redesign at our second regional.
There are tons of benefits! Over the seasons, the teams I’ve been on have had practice machines in 2009 (full), 2010 (full), 2013 (partial), 2014 (full), 2015 (full), and 2016 (partial) and they were instrumental in allowing our programmers, drivers, and mechanisms to develop.
Main Pros:
*Drive practice time!
*More time to improve robot’s program.
*You can easily modify and check subsystems after the deadline.
Cons:
Time and money.