Drivability improvements

[berkakinci]

Add deadband for heading control. Improve fine motion of robot. (With acceleration limit or more aggressive stick profile from power of 3 to power of 5) After competition drivability review with drivers/operator.

[berkakinci]

Pull Request #17 improves fine motion of robot for both AbsoluteDrive and AbsoluteDriveAdv. Pull Request #16 increases heading deadband to 75% (from 50%) only for AbsoluteDrive.

[berkakinci]

I want to close this, but there is still one more improvement we could make: the analog stick profile from power of 3 to 5 or even 7. Have Renuka try both and see which works best.

diff --git a/src/main/java/frc/robot/subsystems/swervedrive/SwerveSubsystem.java b/src/main/java/frc/robot/subsystems/swervedrive/SwerveSubsystem.java
index 6429b2c..a67223b 100644
--- a/src/main/java/frc/robot/subsystems/swervedrive/SwerveSubsystem.java
+++ b/src/main/java/frc/robot/subsystems/swervedrive/SwerveSubsystem.java
@@ -298,8 +298,8 @@ public class SwerveSubsystem extends SubsystemBase {
    * @return {@link ChassisSpeeds} which can be sent to th Swerve Drive.
    */
   public ChassisSpeeds getTargetSpeeds(double xInput, double yInput, double headingX, double headingY) {
-    xInput = Math.pow(xInput, 3);
-    yInput = Math.pow(yInput, 3);
+    xInput = Math.pow(xInput, 5);
+    yInput = Math.pow(yInput, 5);
     return swerveDrive.swerveController.getTargetSpeeds(xInput, yInput, headingX, headingY, getHeading().getRadians(), maximumSpeed);  
   }

BTW, you may want to reduce deadband too. Deadband of something like 2x or 3x noise in stick should be plenty especially since the higher powers will reduce the effect dramatically.

[berkakinci]

I also noticed there are two implementations of getTargetSpeeds, so update both. Second doesn't seem to be currently used. If you make a Constants.java entry for the power, you can make it more generic with a little more math. This one without complex numbers: y = sign(x) * abs(x) ^ a https://www.desmos.com/calculator/uzelyagtci

[berkakinci]

Pull request #21 addresses the final point. I didn't change current behavior, but wrote the code to make it easily adjustable. Tweak away at next opportunity to practice.