PathPlannerTrajectory.transformStateForAlliance appears to be incorrect?

[vfurman-gh]

Hello,
First of all, thanks a lot for sharing the path planner code! It is a big help for the team I mentor!
We encountered an issue with the library, though. We trying to use PathPlannerTrajectory.transformStateForAlliance, looking at the PathPlannerTrajectory.java:124, it looks like translation happens by mirroring around axis with Y=FIELD_WIDTH_METERS/2:
new Translation2d(state.poseMeters.getX(), FIELD_WIDTH_METERS - state.poseMeters.getY());

I think to transform to red alliance coordinates, we should actually mirror around axis with X=8, so the code should look like:
new Translation2d(16 - state.poseMeters.getX(), state.poseMeters.getY());
What do you think?

[hadley31]

The origin is always assumed to be the right corner of your current alliance station. This years game makes that a little more difficult because the field is not symmetrical in that way, thus mirroring over the Y axis is correct

[vfurman-gh]

On Thu, Feb 9, 2023, 4:15 AM Nick Hadley ***@***.***> wrote: The origin is always assumed to be the right corner of your *current alliance station*. This years game makes that a little more difficult because the field is not symmetrical in that way, thus mirroring over the Y axis is correct.

Ok, that starts making sense. So April tags locations need to be transformed for alliance as well? I wonder if it would be simpler to have origin fixed in the blue alliance corner always and have transformation only for PP trajectory? Wdyt?

…

— Reply to this email directly, view it on GitHub <#241 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ASTB463DCTD775RUIR2A4SLWWTNXZANCNFSM6AAAAAAUWDLPX4> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[hadley31]

Yes, the AprilTagFieldLayout class has a setOrigin() method, that should be set depending on the value of DriverStation.getAlliance()