Provide an Example of a Timed Multi-Step Non-Command based Autonomous

[jasondaming]

I think there is a strong case that this should be added to the Simple Autonomous Example in ZTR. Thoughts @Daltz333 / @Kevin-OConnor ? If not where would you put it instead?

Here is the simplest I can think of:

// Auto Time Contants
private Timer timer;
// Modify the numbers on the steps below to control how long each step lasts
private double step1Time = 2;
private double step2Time = step1Time + 3;
private double step3Time = step2Time + 3;
private double step4Time = step3Time + 4;

@Override
public void autonomousInit() {
  timer = new Timer();
  timer.start();
}
    
// Only one of the statements will run
if (timer.get() <= step1Time) {
  // Do nothing so we don't get in someone else's way.
} else if (timer.get() <= step2Time) {
  // Run the elevator and the shooter
  m_elevator.set(elevatorSpeed);
  m_shooter.set(shooterSpeed);
} else if (timer.get() <= step3Time) {
  // Stop the elevator and the shooter.  Drive backwards
  m_elevator.set(0);
  m_shooter.set(0);
  m_robotDrive.arcadeDrive(-0.25, 0);
} else if (timer.get() <= step4Time) {
  // Stop the drivetrain
  m_robotDrive.arcadeDrive(0, 0);
}

[Kevin-OConnor]

The example in there is what is provided in the example project.

I would argue that this goes nowhere. I don't think that a complex multi-step auto based completely on time is a pattern that should be encouraged.

[jasondaming]

So what would you do instead encoder counts? The team I am helping wants to dump into the low goal then back up, that is super simple. I see no reason they should be forced to add encoders to do that process. I agree that sensors make more complex processes easier especially when distance precision is needed but that isn't easy to add for many low resource teams and adds significant complication.

I think we currently have a gap on doing anything more than a single movement in auto and I think this helps change that. What would you propose instead to fill that gap?

[jasondaming]

Also a timed multistep auto is already in the base code for LabVIEW. I agree that I wish more teams would progress beyond that style of auto. But I think it can be successful for teams doing simple things.