The path is known. The design team presented their work, the prototyping teams reported on the lessons learned, we discussed the trade-offs we're facing. The path is known. We don't make decisions, we set direction. Direction is meant to be adjusted as we learn more. The path is known...
The Day 15 Huddle-Up
On Saturday January 20th (Day 15 of our 2018 Robot Build Season) the Robowranglers sat down for what would be a 3-hour long "Huddle-Up" meeting.
The topic of the day? Robot trade-offs.
The stakes? Decisions. Decisions. Decisions about what our 2018 robot will do.
During our prototyping we ran into conflicts. The Robowrangler students began to understand that in a design challenge like this, the constraints matter. The constraints force us to make decisions. The constraints force us to understand what is MOST important to us.
The robot cannot do everything. This is one of the hardest lessons to learn. "At some point in the season, every one of you will try to convince me that we don't need to choose a path and that we can take both paths. We must always choose a path, even if it's a middle-path. You can't walk two paths."
Even veteran engineers sometimes struggle to remember that truth: the robot cannot do everything. By definition, if we focus on one area of performance and truly optimize it, another area of performance must suffer.
Thankfully, while 148's robots only come in one color, the world is not just black and white. We have a spectrum of choices in front of us. We don't choose A over B, we choose where on the line between A and B we want to fall. We choose our desired balance. The robot cannot do everything, but it can do a balance of things.
To help understand some of these trade-offs, and to help us decide what we find most important, we discussed a series of hypothetical robot functions.
This list includes various things the Robowrangler robot might do, and for the purposes of the exercise we pretended we already decided to accomplish all these functions, then - we did an old fashioned pro/con list.
(Reminder for "our viewers at home" this is a thought-exercise, not a list of Robowrangler 2018 Robot Functions... I don't think people would be happy if I published that list.)
Our robot will be able to hit the corner of the ramp at full speed and do "sick jumps" without sustaining any damage!
Our robot will be able to climb ANYONE's ramps.
We will be able to climb from the bar while helping a large partner robot climb.
We will be able to "limbo" drive under the scale, even when it is tipped down in our favor (which we hope it will be most of the match).
We will intake off the front of the robot, and score off either the front or back of the robot.
We will be able to grab cubes directly from the human player in the portal.
We will put a "dump truck" (Everybot) mechanism on the back of our robot.
What are the pros and cons of each of these hypothetical decisions?
Based on the lists of these pros and cons... how do we feel about these functions?
Of course, this list isn't ALL the things the robot could do, these were chosen for one purpose: "to make Wranglers think".
The designers talked about the impact each decision would have on the system design and overall performance. They showcased some of the "foundation" work they'd done over the past week. They demonstrated how things would fit together in each of these hypothetical black robots.
The prototypers talked about what they'd seen in their experiments, and provided some "real world" data to support (or refute) the claims of the designers. They helped us understand the "effort" side of the "effort vs reward" trade-off.
The strategists talked about how these functions would impact our ability to fulfill our Goal Map. How the robot would win matches with or without these features. They helped us understand the "reward" side of the "effort vs reward" trade-off.
This 3-hour marathon huddle yielded... a plan. We have a list of what we value. We have a direction. The path is known.
The huddle-up ended with purpose. The prototypers go back to their experiments. The strategists back to their simulations. And everyone knows what the robot will look like. The path is known. The design team went back to work with the familiar mandate of each year. "Deliver the CAD. Deliver it now. Make it great."
The official meeting ended at 10PM, but the design marathon was just beginning. Systems Integration is the order of the day. With a green-light from the team the designers are running towards the goal line. But, around 11PM we discovered a problem: It doesn't all fit.
The Design Push
"How do you cram 10 lbs of robot into a 5 lb box?"
11:00PM Saturday - We discover there isn't as much room in the back of the robot as we expected.
11:30PM Saturday - Decide to rearrange things. Lift subsystem moves forward on the robot. Everyone pats themselves on the back since we have plenty of room for it and the intake actually gets smaller and lighter in this setup. Lift gearbox needs to be redesigned. Drivetrain needs some tweaks to accommodate the new position. Team splits up to do some integration CAD work for each subsystem.
1:00AM Sunday - Things are falling together. CG is discussed a lot. Everyone agrees this new setup is much better. "Why was the intake so far back anyway? There was all that dead space behind the hard-stop."
2:00AM Sunday - Most people go home.
5:00AM Sunday - The last die-hards call it a night. The robot is "all there" and detail work is moving forward in earnest on all subsystems.
11:00AM Sunday - The "early" crew wanders in and discovers that someone from the 5AM crew is sitting on the -000 assembly so no one can save changes to the overall system.
2:00PM Sunday - The rest of the design team arrives and detail work on the robot begins with renewed fury. The person who was holding ownership of the top-level CAD model is shamed, but ultimately forgiven in the spirit of "late night CAD camaraderie".
4:00PM Sunday - The robot is nearing "complete integration" and people are starting to do lightening patterns and final mounting holes.
4:05PM Sunday - Discussion pops up about whether the proper nonagon placement is "point up" or "point down". (The correct version is point down, wings up. Always.)
6:00PM Sunday - The design team talks about moving the winch gearbox to a new location. Art Dutra just stares at them until they quietly go back to work. The winch gearbox does not move.
7:00PM Sunday - We break for burritos. Mmmmm.
8:20PM Sunday - Madison has a realization that the intake geometry has a flaw in it which the design team didn't notice at 11:30PM the night before (or in the hours since).
8:20PM to 8:30PM Sunday - Madison triple checks her realization and compares it with the prototype measurements.
8:30PM Sunday - Madison tells the rest of the design team we have a problem. Cue the 7-stages of design grief.
8:31PM Sunday - Shock and Denial - "It totally isn't broken. We checked it."
8:32PM Sunday - Pain and Guilt - "It can't be broken. We'd need to redo 24 hours of work? JVN aren't you supposed to check stuff like this?"
8:33PM to 8:49PM Sunday - Anger and Bargaining - <insert profanity here> "What if we move it like this? Does it work then? No. Okay, what if we just add another mechanism which flips out a thing to fix the problem?... ... No, that's not stupid, you're stupid." <insert profanity here>
8:50PM Sunday - Depression, Reflection, and Loneliness - "... ... ..."
8:51PM Sunday - The Upward Turn - "This is for the best. Most of our work is still good. We can make it better anyway. with this new change. It's an important change."
8:53PM Sunday - Reconstruction & Working Through - "Okay, so we need to redesign. Here's what we're gonna do... move the lift back a bit, move the battery back up to the front of the robot. This makes the CG happier - we should have done that anyways. We'll redesign the lift gearbox again. We'll extend the drivetrain by a few inches, and it should all work again - right?"
9:00PM Sunday - "Alight, everyone go home. We'll pick it up tomorrow night. I'll let the team know about the issue. Design is iterative..."
Welcome to Robowrangler Design. Fail. Learn. Fail Again.
"We don't make decisions, we set direction. Direction is meant to be adjusted as we learn more."
The path is known.
Knowing the path is sometimes the hard part, but walking the path ain't no picnic either...