Paths different when deploying from different computers

When we deploy code from our competition’s laptop the robot follows the path correctly but when the same exact path is uploaded from our main work desktop the path is around 45 degrees offset. Does anyone have an explanation for this very weird behavior? I know for a fact it’s the same path because I pushed the changed code to our GitHub and then pulled it to the comp laptop so I know the code is exactly the same.

Check your computer’s clock and make sure it is correct. We had a problem with falcons when our clock was incorrect

1 Like

Alright. I just checked the time in both to make sure they are exactly the same and they are, so that is not the issue. Any other ideas?

Maybe make sure you pulled the right branch/commit?

Check on both machines that you don’t have any uncommitted files.

I just checked to make sure origin was set to the same thing on both of them and both were set to the main branch on the correct origin.

I just checked that as well and no files are untracked or not committed. I am so very confused how the same exact path code on both of the computers are producing different results.

Does the robot odometry (e.g. using Field2d) think that the robot is following the correct path?

The path appears to be visibly different between the two laptops.

Could code somewhere else be different? E.g. constants that define max acceleration.

I’m not sure exactly what you mean here. When you say “path”, do you mean the physical path followed, the planned trajectory, or the measured odometry?

Can you post a screenshot from each of the two systems showing a Field2d with both the trajectory and the odometry?

I have not used Glass before and am very new to the whole Trajectory thing anyway so bear with me. This may not be what you are looking for but I think it is. I am not able to work with the robot the rest of tonight so next response may be delayed.
Desktop:
Desktop
Laptop:
laptop

I know the page I linked to refers to Glass, but the same robot code should work equally well on ShuffleBoard. Sorry for the confusion; I should have pointed that out.

If I’m reading these screenshots correctly, the robot is being asked to follow the same trajectory, and it believes it is doing so. (Apparently the Glass defaults make for a messy display.)

Are you triggering the path in exactly the same way on the two computers? Can we see your code?

Maybe make sure all packages are updated on both computers?

Try checking compiler, VS code, and vendordeps.

Are they both windows machines?

Very sorry we have been very busy with the comp so I was not able to get a response till now. Both computers have now developed the same issue so computer differences are off the table. Our most recent code is here. We are sadly done with this season but I would like to know what the issue is for next season.

This topic was automatically closed 365 days after the last reply. New replies are no longer allowed.