Concept Lock

[TEE]

I was looking at one of the simbotics online presentations (http://www.simbotics.org/resources/workshops), and I saw that they reach concept lock at the end of week 1. This year, my team tried locking the concept, but we had to change things because our original concept didn't really work (and that concept lock was halfway through week 2 or 3).

Now I'm wondering: how on earth do you strategize for the game, and conceptualize a full robot in a single week

[DonRotolo]

You get better with practice. Don't expect to get to 1114's level in your first try, but after a few tries it should happen. The idea (trick?) is to learn the process well and be very productive in that week - recognizing and discarding unworkable concepts early while keeping creativity and brainstorming alive.

I like to think that after 5 years of trying our darndest we're at the point where excellence is within striking distance. Even though we won 2 regionals this year, we're not at the 1114 level. Yet...

[Chris is me]

This year it was pretty hard because the first 4 days of the week we were all building different robots, and after that I think all of us changed our designs significantly.

[J_Miles]

As a team, we're attempting to get to that point. A lot of it comes from experience. Having mentors who have been around FIRST since long before our team started, as well as having students that have dove into research from past years and really gotten a lot out of their three or four years with the team has eased the process significantly

The trick is deciding first HOW to play the game, independent of any sort of machines or mechanisms, within the first day or two of kickoff, and then designing a robot to play the game the way you've decided.

By doing this the different robot concepts are far more focused, and it becomes a matter of deciding which design does what you want best. You don't have to decide, then, which concept is going to achieve what you want, and then further refine that concept to do what you want WELL.

Being only a fourth-year team, we struggle with this, as well, but the turnover between kickoff and "concept lock" has deflated dramatically each year, and, I think, soon one week will not be at all unreasonable!

[Lil' Lavery]

While many of the best teams arrive at a concept very early in the build season, there's more than one way to skin a cat. There are elite teams that don't start building their final robot until well into the season because they're still prototyping various concepts.

The common denominator either way? They work their freaking butt off. Crunching in all that design process into one week takes a lot of work, as does assembling and debugging a robot in the last two weeks of the build season. Go ask the elite teams what type of hours they put in during build season.

[Peyton Yeung]

Team 45 has been around since the beginning and we still struggle with design "locks". Typically we spend about two-three days in subgroups to come up with concept ideas which are presented and the team votes on design ideas that are considered the best. Of course 1/2 through the season we decide the original design of certain components aren't efficient so we change them. For example, our gripper started as a roller claw. We found this design didn't work well so we changed it into a pincer claw. After more designing we have evolved our claw to work better than any other claw we have designed for this game. "Locks" on design are necessary but our team drives itself by believing that there is something that can be done to make our robot better.

[tim-tim]

I think, as some have said, it is way more important to develop a strategy lock. This will help you dictate the rest of the season.

Not knowing exactly what 1114 does during the build season, I cannot speak directly towards their process. I can however speak on ours; we sit down and develop a strategy of how we want to play the game. This process usually takes most of Saturday and then Sunday...sometimes into Monday. At this point we start the "How are we going to complete this strategy?" questions.

Working in industry for a brief period of time really showed the importance of iterative design. Getting feedback from the customer, shop technicians, etc is a must and really allows the design to become better. How do you implement this is FIRST? Do your research, check CD, Prototype, and prototype.

To sum it up:
- Develop a "Strategy Lock"
- Allow actual robot design to become an iterative process with the "Strategy Lock" being a check to make sure you are on the correct path.

[J_Miles]

Originally Posted by tim-tim

To sum it up:
- Develop a "Strategy Lock"
- Allow actual robot design to become an iterative process with the "Strategy Lock" being a check to make sure you are on the correct path.

This is very similar to what we do. Strategy comes first, and strategy dictates our design. We have a concept lock, essentially, but that doesn't mean that we can't engage in an iterative design process. The progress of our minibot throughout the season is proof of this. We're constantly aiming to improve, and allow for that; however, establishing concrete goals early in the build season allows the six week build to be focused towards a single point so that disaster does not strike and the robot may actually be close to completion BEFORE ship day.

[DUCKIE]

Don't know if this will help... and MOEmentum probably has better information, but MOE starts off with the following basic process every year...

Day 1: Watch Kickoff, Re-watch game animation a couple times, break down scoring/field/rules basics while also creating a question-list for mentors at kickoff/remote kickoff locations or things/field parts to get pictures of. Eat Lunch. Divide up into 4-5 general strategies -like aggressive defense or offense- Student homework is to read the rules and be ready for tomorrow.

Day 2: Parent volunteers and couple mentors oversee the construction/inflation of important field elements / scoring items while students take part in brainstorming. Students encouraged to draw/sketch anything and everything... from an autonomous program track or gripper design to a drive train idea or an entire robot. While students and mentors take turns presenting their ideas to the group, the ideas are taped up to the wall in groups with similar concepts/robot parts. Usually before or after a short visit to see the under-construction field parts / prepared scoring items, each person is given so many vote ticks or dot-stickers to vote for their favorites ideas. Top items or concepts are collected after voting, and are used for the designs as well as to how sub-teams will be broken down for the build season.

From there we are able to finalize a robot concept or at least strongly narrow down what will be built with everyone before the first week is up... although some items [like last year's kicker or the bin lifter in '04] need mock-ups and testing by relevant sub-teams to finalize.

[The Lucas]

Originally Posted by DUCKIE

From there we are able to finalize a robot concept or at least strongly narrow down what will be built with everyone before the first week is up... although some items [like last year's kicker or the bin lifter in '04] need mock-ups and testing by relevant sub-teams to finalize.

The #1 thing that "throws a wrench" in achieving concept lock in Week 1 is rules clarification and the slow turn around time of the Q&A. It take a significant "leap of faith" to lock a concept (primary strategy) when there is a significant chance that that strategy will be declared illegal in the following weeks.

For example, last year by Day 2, we were dropping soccer balls onto the bumps and noticing we could score them into or get close to the corner goal. In the next few days we built prototype plywood ramps to demonstrate this concept (crude but a similar concept to 469's amazing machine). Our strategy was to drive onto the platform using the bump in auto. We built a prototype 4wd drivebase with a pop-down turntable in the middle to execute a 90 degree turn on top of the bump. By climbing onto the platform it was hanging and had a neat idea to suspend partners. We waited longer than usual to lock in a concept (giving more time to demonstrate the redirector concept), but when it finally came time to vote between this redirector and a variable power kicker we still didn't know if the redirector would be legal (it looked illegal by the kickoff rules). The variable kicker won in a land slide and I was the only person to vote for the redirector gamble. The 4wd on the competition bot is an artifact of the prototype for the redirector bot and time not spent prototyping an 8wd. the Kudos to 469 for rolling the dice on their strategy that was later clarified t be legal. We have hit snake eyes on those rolls in the past so we are averse to those big risk, big reward gambles on unclear rules.

It is important to get everyone on the same page and move forward as a cohesive group once the concept is locked. If you have a dissenting opinion make your case before the concept lock. If your concept isn't the choice of the group let it go and focus on making the locked concept the best you can in the few remaining weeks. It may not be what you consider the "best concept", but all concept have pros and cons like all engineering decisions.

If you can't tell by now I have significant experience being the dissenting opinion to at least one aspect of our eventual locked concept. This year I made the case for a mecanum drive instead of 6wd and argued that "pushing power is of little to no use for a tube scoring bot in this game. In fact, pushing a defender into your zone is one of the worst things you can do since they would be in your zone blocking your rack without a penalty under G61". I didn't win this argument, and I ended up being very wrong in Wk4 of build after TU6. TU6 added G61 exceptions, dramatically changing the game and undermining my argument (pushing power was now quite valuable).

In summary, my advice is
-Lock a concept early, usually around week 1
-Even if you don't fully agree with the concept, work together to make that concept the best it can be.
-Anticipate potential rules changes in Team Updates and appropriate changes to your concept (particularly if your concept is "risky"). However, don't waste resources developing these changes until they are necessary.
-Sometimes your concept will work well and sometimes not. Sometimes after weeks of work, you find out you can't build a mechanism to implement part of the concept satisfactorily within the many constraints. Even though you may fail, it is still better to focus on one concept the best you can. The rules may be a bit of a moving target with all the Updates, but there is no tougher moving target than multiple ever-changing concepts.

Our robot this year is very much like the concept we put forward in Week 1. The big change is the PVC surgical tubing slingshot minibot had to be completely scraped due to rule changes. We already had a great prototype for that minibot, but at least that rule change was early (TU1 in 3 days).

[Hawiian Cadder]

we design something very flexible. this years robot was designed with multiple redundancies in mind.

the slider drive became 6 wheel tank

the lift is set up in such a way that it could be replaced with a 4 bar or arm. the manipulator was designed in such a way that it could be mounted to an arm

the manipulator could be replaced with a simple clamper.


for next year i layed out a simple frame that is essentialy an upgrade from the KoP. using configurations, and by altering a few linear patterns it achieves an outstanding level of flexibility.

it can accommodate any drivtrain, mechanum, holonomic, crab, up to 10 wheel tank. the drive base is also set up to support both internal manipulators (think kickers from 2010, and conveyor belts from 2009) as well as external manipulators (arms, lifts, anything that does not take up a large amount of space in the belly of the robot.

the aluminum parts of the frame weigh only 9 lbs, and for every wheel you add .6 lbs.

crab modules were not cadded, because of the extreme unlikelihood that we would suffer the weight and complexity.

[LightWaves1636]

On our team we prototype A LOT in the off-season. We take our students and study every old FRC game and start CAD work and 2D's for end effectors, mechanisms, and chassis designs. If we have the resources, we'll actually build the robot.

By the time Build Season rolls around, we lock down what we(The Mechanical Sub-Team) want by the end of Sunday. Chassis 2D's is due at the end of Sunday. Due to our preparation and anticipation, Build Season just becomes about integration.

The more you do off-season as a full team, the better and easier build season will run.

Of course there are some few exceptions where we haven't prepared for like minibot.