Auto creation work

[Hsaunders603]

Autos to make:

• [ ] [2] Score preload -> charge station (14-18)
• [ ] [3] Score preload -> out of community (pickup?) (9)
• [ ] [2] Score preload -> charge station -> pickup (out of community)-> charge station (17-21)
• [ ] [1 or 3] Score preload -> pickup (out of community)-> place piece (15)

  ---

  [1] = starting position near barrier [2] = starting position in middle (in front of charge station) [3] = starting position side near judging table (wire run)

  ---

[TimB-87]

We're back!!

New Project: Auto that drives a swerve drive-base in a square (between four points drawing out a square)

Potential ways to do it:

• Use pathing the correct way (really unsure how to do this as I could not figure out PathPlanner last year)
• Just tell the robot to move this fast for this long to end up going this distance one way then another then another, etc.

[msoucy]

Test work is done in #75

[TimB-87]

Notes/Thoughts

(After testing the PathPlanner square auto on 11/14 and conducting research into the auto of other teams and our own from 2022)

Waypoints, waypoints, WAYPOINTS

⬆️⬆️⬆️⬆️⬆️⬆️ *Pretty sure this is what we've been missing

---

In contrast to PathPlanner, there is PathWeaver, which some teams use, it seems like PathPlanner will work just fine if not better though.

---

Team 1678 seems to use PathWeaver, along with using modes and actions rather than sequences.

• https://github.com/frc1678/C2023-Public/tree/main/PathWeaver
• https://github.com/frc1678/C2023-Public/tree/main/src/main/java/com/team1678/frc2023/auto

---

Team 4499 has a pathing tool, though I'm not sure how exactly it works. It seems like they set it up to click on their computer screen and layout a path.

• https://github.com/HighlandersFRC/2023-FTC/blob/main/FtcRobotController-8.2/TeamCode/src/main/java/org/firstinspires/ftc/teamcode/Autonomous.java
• https://github.com/HighlandersFRC/2023-PathingTool/blob/dev-Chezy/pathtool.py)

---

Team 6328 has a lot, but they use waypoints along with Pose2d, so it could be worth more looking into

• https://github.com/Mechanical-Advantage/RobotCode2023/tree/main/src/main/java/org/littletonrobotics/frc2023

---

In the past (2022), we have used PathPlanner and waypoints and it seems to have worked with PathPlanner.loadPath, we might not need Pose2d for the auto paths

• https://github.com/chopshop-166/frc-2022/tree/main/src/main/deploy/pathplanner
• https://github.com/chopshop-166/frc-2022/blob/main/src/main/java/frc/robot/AutoPaths.java

Overall, lets see about doing exactly what we did in 2022 and use waypoints along with PathPlanner. We might still be able to use Pose2d though, it would just require a deeper look into 6328's code.

[TimB-87]

Notes / Updates

New Findings: Poses are very similar to waypoints, and they run on a coordinate system. Last year, we did not get the correct coordinate system and the robot was trying to go to the coordinates we plugged in, but those were the wrong coordinates -- Poses are what we need

Meeting of 11/16/23

End of meeting, we figured out how exactly to use the coordinates for poses. Using PathPlanner with Poses, we have an auto that will have the robot drive in a

• square (left, then forward, then right, then back) facing one direction, and one for a

https://github.com/chopshop-166/frc-2023/assets/119461590/6ec8c4bc-5320-434c-93a5-8bbfd463876e

• triangle

https://github.com/chopshop-166/frc-2023/assets/119461590/a1e2481d-c400-4d21-baaa-600496dc57d3

• Both have wait commands in between points.

Both of these autos are quite shaky, though I believe it is just because of the drive function. At the current moment, PathPlanner uses the drive function to drive the robot from one place to another. The robot speeds up just enough and then slows down to stop at a point.

Meeting of 11/21/23

First things first, we removed the wait commands in the square and triangle autos to speed them up. They worked great without the wait commands, though the robot sometimes goes a little past where it is supposed to stop and ends up thinking it's at a different coordinate then it is

We created a line auto from one side of the carpet straight to the other to test how fast the robot goes. We came to the conclusion that the further the robot drives in auto, the faster it goes to get there because it has more time to speed up.

Next, we played with rotation a bit. It seems that from one point at rotation 0 degrees, to another point at rotation 180 degrees, the robot will turn on the way and end up fully turned however much it was told to at the point.

https://github.com/chopshop-166/frc-2023/assets/119461590/31deab3b-9ed6-4673-92c9-c32b6486cc66

With this, Layth suggested a "game" where the robot would drive straight, drive to the right while turning that way, and knock one of the carpet rollers off of the carpet, then return to the starting point. This was very successful, so we came up with an auto that would do it twice, affectionately labeled the "Knockout" auto.

https://github.com/chopshop-166/frc-2023/assets/119461590/8df211a3-9c56-4e19-8779-e33cdc280fcc

Main Problems

• Coordinate Reset: Only current way to reset coordinates (which we constantly need to do to get the robot to think it's at the right (0,0) coordinates) is by turning the robot off and on again or putting it on the robot cart and running the auto. I hope to make a function similar to the reset gyro function that will fix this.
• Shakiness in Driving: The driving in the auto is very shaky, though I don't believe it's the fault of the auto. I plan to adjust the PID constants here: https://github.com/chopshop-166/frc-2023/blob/main/src/main/java/frc/robot/maps/FrostBiteMap.java#L119
• Speed, potentially: From point to point, the robot speeds up and then slows down to stop at the next point. I'm not sure if there is a way to speed the auto up because of how drive works, but there may be another way to set it up to accomplish the same goal.

For now, the goal is accuracy, then we will focus on speed

I am having a lot of fun with auto, so we will be doing a meeting on 11/22/23 to hopefully fix some of the main problems and create an auto involving picking up and transporting a cube.

[TimB-87]

Notes / Updates

Meeting of 11/22/23

Unfortunately, some weird difficulties came up, and the drive was being very terrible for about 20-30 minutes. Luckily, it randomly went away.

Tonight, we mainly worked on

1. A better way to reset the robot's coordinates.

   • I worked on trying to make a function for a bit, and I could not figure it out. Though, I found out that just resetting the robot code on the drive station laptop does the trick and only takes about 5 seconds!

2. An auto that picks up a cube, returns to the "community" and then deposits the cube

https://github.com/chopshop-166/frc-2023/assets/119461590/57b4c0af-2dd7-4df7-b08c-1dda5e9c0349

Next Steps (in order)

• Tune drive PID
• Speed up this auto a bit (including seeing if we can speed up driving) - Get us to 15 seconds!
• Potentially, as a goal for "do by build season", do this same auto with Vision

[TimB-87]

Notes / Updates 12/15/23

Above is the makeshift auto. It is our auto to fall back on if everything goes wrong, but hopefully we'll never need to use it.

We moved on to working on PathPlanner 2023 for the past couple weeks, as it seemed that would speed things up quite a bit. However, we did not get to testing it as we have not finished yet. Then, the code was updated for 2024.

Moving on to PathPlanner 2024

We have decided it is in the best interest to just move on to PathPlanner 2024 and figure it out as soon as possible. We've experimented with writing PathPlanner 2023 sufficiently enough that we are familiar with the system. However, PathPlanner 2024 takes a lot of PP 2023 and puts it in the library, making our lives a lot easier.

We hope to have running autos with PP 2024 before the build season.

*Note: in terms of vision, we will be focusing on auto first, and after that my hope is to have a simplistic vision working on the robot in the 2024 season.