/2022-2706-SwerveSummer

Doing swerve in the summer.

Primary LanguageJavaOtherNOASSERTION

2022-2706-SwerveSummer

Doing swerve in the summer.

This codebase originated from the WPILib example code avaible here: https://github.com/wpilibsuite/allwpilib/tree/main/wpilibjExamples/src/main/java/edu/wpi/first/wpilibj/examples/swervecontrollercommand

We removed the PWM motor controlers in the WPILib codebase and replaced them with CAN motor controllers.

Currently supports SDS mk4.

Also supports Swerge V4 with SwerveModuleSparkMax however the DriveSubsystem needs changes to use SwerveModuleSparkMax.

Going to Kingston to test code

Thank you to team 7480 Machine Mavericks once again.

Lessons Learned / Interesting things to note

  • SwerveControllerCommand is awesome. It uses the trajectory objects to control the X and Y then it controls the heading of the robot separately.
  • We had issues controlling the heading of the robot in auto. I think the gyro was not rotating the correct direction for SwerveControllerCommand but I'm not sure (it needed to be multiplied by -1 I think)
  • We took advantage of the gyro offset inside SwerveDriveOdometry for Teleop. Since we also set odometry to the start of an auto path it means we can start the robot at any angle and when teleop starts it's already correct.
  • The code given by SDS doesn't run a PID loop on the velocity of the wheel. I think this is important for SwerveControllerCommand because when it asks kinematics to go at 2 meters per second it expects it to do it perfectly (and adjust the position later if it's not perfect)
  • We only tuned the F and P of the velocity PID loop. It seemed to work just fine with only these 2 terms (it was achieving the desired velocity with only a tiny degree of error).
  • SDS code changes the status frame of Status_1_General to 250 ms of all 8 falcons running the modules. Status frames, from what I understand, affect how often the CAN bus gets updated with specific values. Status_1_General says it's the "Feedback for selected sensor on primary PID[0]." I was scared this would affect odometry with outdated wheel velocities and steering angles so we didn't add this to our code.

Teleop

After tuning the PIDF and IZone values for the drive and steering motors, kinematics worked great to control the robot in teleop.

Video of teleop control: https://www.youtube.com/watch?v=8znFHGSqZx4

Autonomous

Once Kinematics is tuned with a PIDF for velocity control and PID for steering, autonomous came shortly after.

Autonomous uses the SwerveControllerCommand provided by WPILib: https://github.com/wpilibsuite/allwpilib/blob/main/wpilibNewCommands/src/main/java/edu/wpi/first/wpilibj2/command/SwerveControllerCommand.java

This command uses the Trajectory objects to control the X and Y of the robot. It also has different constructors to state what the heading should do while it goes through the X and Y of the trajectory.

SwerveCommandMerge is something we created to make it easier to construct SwerveControllerCommands. SwerveCommandMerge only needs to be given a trajectory and what the heading should do instead of the 7 parameter long constructor that SwerveControllerCommand needs.

Autonomous test1 - Slow Speed

Video of slow speed: https://youtu.be/DfByJQQnnsY

First thing to note. We disabled the SwerveControllerCommand from controlling the rotation so what's seen in the tests is the heading drifting and not correcting itself. We had issues with the rotation and not enough time to investiage them, so the video is only demonstrating the X and Y control.

Path planned: PathShownInVideo

Shuffle board reporting the error in the X and Y of the path. The veritcal axis is in meters. AutoPath-SlowSpeed-Shuffleboard

Autonomous test2 - Fast Speed

Same path as slow speed.

Video of fast speed: https://youtu.be/vVuHyKpA6Yo

Shuffle board reporting the error in the X and Y of the path. The veritcal axis is in meters. AutoPath-FastSpeed-Shuffleboard

Instructions

Part 1 - June 6th

Started with a 20 minute presentation about what is Swerve.
https://docs.google.com/presentation/d/1UlFns79VPLa6JWTARuPIhunhoWoVWEZtnUoII4OY9Xs/edit?usp=sharing

The last 2 slides have some instructions which are copied below:

  1. Clone SwerveSummer2022 and make a branch with your name as the title. Your branch should be from the NewStudentsStartHere branch.
  2. Setup 2 CANSparkMax objects in SwerveModule.java
    - Set PIDF+IZone to 0 for now. We will tune them on another day.
  3. Complete setDesiredState method and getSteeringAngle method in SwerveModule.java

Check out this PDF for a guide to which CANSparkMax settings need to be set.
https://drive.google.com/file/d/1IVFa-3UEYdORYtp_CFTkWB-YxjUFXocq/view?usp=sharing