Kickoff Retrospective
Written by @MOCTA1189, @Nova-Gear, and @Gavin0
Welcome to REEFSCAPE! Yesterday was Kickoff, so we’ll be using this post to inform you of our strategic and design processes, as well as some of the early concepts and prototypes that began forming throughout the day.
Our Approach and Process
After returning from the In-person Kickoff Event hosted by the University of Detroit Mercy, the first hour or so was a rapid flurry of pizza, reading the game manual, and discussing everything any of us noticed and our first impressions.
From there, we transitioned to the second phase of the process; determining what to prioritize.
We created a Dreams | Wants | Needs table to decide what capabilities we need at minimum in order to optimize our planning and design to our capabilities. Although there is some degree of strategy involved in determining this, our approach this year has shifted toward predicting strategy enough to know what capabilities are valuable to have, without being bogged down by getting too granular with it.
Our expectation is that there are a few basic areas that will be strategically advantageous to capitalize on, and as long as we can do those basic things, we’ll be able to adapt to a meta that we will have the opportunity to observe developing in the 3 weeks before our first competition.
Much of our competitive success over the last few years has come from the relative simplicity of our robot designs, so being able to focus on what we really need to be able to accomplish is a key focus of these early stages.
Our determinations have been listed below in a table for clarity.
| DREAMS |
WANTS |
NEEDS |
| Buddy Climb |
L4 in Auto |
Score CORAL at all levels |
| Shoot ALGAE into NET |
2-3 CORAL Auto |
Remove ALGAE from REEF |
|
Move ALGAE in Auto |
Drive |
|
Help push Alliance Partner in Auto (to achieve AUTO RP) |
Climb deep CAGE |
|
Climb shallow CAGE (additionally) |
Durable |
|
Floor CORAL intake |
Score ALGAE in the PROCESSOR |
|
|
Score at least 1 CORAL in auto |
|
|
Intake CORAL from CORAL STATION |
From there, we began to split into different groups to work on discussing, rapid-prototyping, and CADing up some concepts for game piece manipulators, robot architectures, and climb.
Idea A: T.O.M. (Total Option Motions)
In our first group, we first decided on two main points.
- One of the most optimal cycles, and definitely something we want to be capable of, would be intaking CORAL, moving to the REEF, removing an ALGAE and keeping possession of it for later scoring or other, and then without moving, being able to score a CORAL where we just cleared up.
- The other point being that a ground intake, at least for CORAL, would be extremely valuable and potentially save a lot of time. During teleop, HPs can drop CORAL while robots aren’t there, and they could roll relatively far into the field, being able to travel less far to pick them up and not wait for a HP to drop them into the robot would greatly reduce cycle time. (And robots are definitely going to be dropping coral trying to score them as well). During auton, even when ignoring the 3 coral marks, using object detection to pick up ALGAE dropped by the human player is a potentially faster option as well.
With these two priorities in mind the group came up with a general architecture of what mechanisms we might need. Starting with a ground intake that can pick up coral from any orientation and store it in a handoff area, as well as potentially intake ALGAE off the ground as well and store them in another area for scoring.
From there, the end effector would pick up CORAL from the handoff point and be moved with some type of arm to the scoring location on the REEF, where it would then be able to drop CORAL. The general idea for the end effector was inspired by the wrist on our 2023 robot, which could rotate up and down and had adaptive FINRAY grippers for holding the game pieces.
We aren’t sure if adaptive grippers were the way to go yet, but we thought about something like this with two different mechanisms on the end effector offset by 90-180 degrees, one for holding and dropping CORAL, and another (either a stick or powered wheels or other) that removes ALGAE and places it into our robot where it can be scores. Ideally this would allow us to remove ALGAE and then score CORAL without ever moving the drivetrain.
Our 2023 Robot
Our group didn’t settle on anything more specific for the mechanisms, but thought about and tested various options for each. We knew we wanted the intake to be touch it own it and be able to center the CORAL, so after testing how the coral interacted with compliant wheels, one of our first ideas was a v-shaped intake of compliant wheels, that could actuate to close into a straight line once the coral was mostly centered.
It is similar to the first design shown by Big Sky Robotics Ri3D team in the linked video.
Coral onto Reef Device (Z-lift) and Cage Climber Group
Our second group brainstormed and designed basic 2D CAD designs for moving coral from an intake to any reef spots and cage climbers. An important consideration we made during design was making strong and fast mechanisms. Faster placement, movement, and object manipulation would likely increase our chances of winning via more cycles. Our group has come up with three REEF ideas–Although, the last one is the most likely to be implemented.
In the group’s opinion, the least likely to work is a telescoping elevator or upward telescoping arm. First, the device would need a wrist on top of the device with at least 180 degrees of freedom for picking and placing CORAL from our intake and onto the REEF. Second, The team believed a telescoping version wouldn’t be able to “pop up” before reaching L4 and the elevator would be too top-heavy leading to tipping. Finally, we couldn’t see an easy option for the wrist to place the CORAL onto the REEF.
The second design was a virtual four-arm lift. The design would have two vertical bars and a simple virtual four-arm lift with a wrist. The idea was inspired by our team’s previous robot from CHARGED UP!, Freddie. We did enjoy the idea of its ability to rotate CORAL easily and, with the right grippers, be done extremely efficiently. However, the arm’s large size and need for volume conflicted with ALGAE. If it had an ALGAE storage, the arm would need to remove the ALGAE to place CORAL– causing wasted time. We determined we needed to drop the design for another possible idea.
Our Final, Current Day 0 plan for the REEF is a “parallel Z-lift”. It is extremely similar to a scissor lift but without the mirrored side. This type of lift keeps its first and third bars parallel and by rotating the whole mechanism we can quickly move the CORAL from one side to another. This mechanism can compress itself into a low vertical profile, lowering the center of mass. In addition, If designed correctly the arm could remove REEF ALGAE and hold ALGAE in place until it is disposed of in the processor. However, testing and prototyping are needed to ensure this design is feasible.
For CAGE climbing, we decided early on that a Deep CAGE has more strategic value than a shallow CAGE. Since a Deep CAGE is so low to the ground we can have a physically smaller mechanism for the work. Currently, we have two ideas: Vertical Drive-in and True Drive-in.
Vertical Drive-in was chronically first and therefore be described first. First, have two clamps–that can rotate vertically freely– on the climbing side of the robot and make those clamps touch the bottom plate of the CAGE. Then, drive forward into the CAGE and let it go into the robot. Using some testing, figure out the rough height and angle of the CAGE once the robot can’t go farther and have extending clamps, hooks, etc. To grab onto the bars and perform a “pull up”.
Side mod: Instead of clamps for the bottom, perhaps use a metal bracket and slot it between the CAGE bars. This allows the bracket to hold all the weight and straighten the CAGE while the upper clamps only ensure the robot doesn’t fall backward.
True Drive-in is how we decided to eliminate the tilt problem of Vertical Drive-in by attaching around the bottom of the CAGE and pulling the cage through the robot. The CAGE is, again, 3.5 inches off the ground. The robot drives into the CAGE until it reaches the Climbing Area. It then clamps, which center the CAGE so the down-facing pistons always have a surface and the center of mass is centered. Then move the robot an inch or two up and True Drive-in is complete.
This is similar to the idea shown here by Penn State Ri3D.
What Went Well?
Overall, we’re satisfied with how Kickoff went. As a whole group, we were able to understand the game relatively quickly, lawyer some interesting new rules, determine priorities, and start working on some ideas.
In addition to that, our fabricators (The Shop Rats, as we call them) started some great progress on developing an at-home solution to the REEF Element. More thorough information on that is coming soon.
What’s Next?
- More Prototyping
- Starting Controls Stuff
- Field Build
Additional Note: Our CAD is now public!
After some delay, our Onshape is now visible for all to observe as we work throughout the season
